Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Lateral entorhinal cortex lesions impair both egocentric and allocentric object–place associations:

Lateral entorhinal cortex lesions impair both egocentric and allocentric object–place associations: During navigation, landmark processing is critical either for generating an allocentric-based cognitive map or in facilitating egocentric-based strategies. Increasing evidence from manipulation and single-unit recording studies has highlighted the role of the entorhinal cortex in processing landmarks. In particular, the lateral (LEC) and medial (MEC) sub-regions of the entorhinal cortex have been shown to attend to proximal and distal landmarks, respectively. Recent studies have identified a further dissociation in cue processing between the LEC and MEC based on spatial frames of reference. Neurons in the LEC preferentially encode egocentric cues while those in the MEC encode allocentric cues. In this study, we assessed the impact of disrupting the LEC on landmark-based spatial memory in both egocentric and allocentric reference frames. Animals that received excitotoxic lesions of the LEC were significantly impaired, relative to controls, on both egocentric and allocentric versions of an object–place association task. Notably, LEC lesioned animals performed at chance on the egocentric version but above chance on the allocentric version. There was no significant difference in performance between the two groups on an object recognition and spatial T-maze task. Taken together, these results indicate that the LEC plays a role in feature integration more broadly and in specifically processing spatial information within an egocentric reference frame. Keywords Hippocampus, spatial memory, associative, episodic memory, navigation, medial entorhinal cortex, cognitive map, landmarks Received: 1 April 2020; accepted: 11 June 2020 2006), border cells (Barry et al., 2006; Solstad et al., 2008), con- Introduction junctive cells (Sargolini et al., 2006), and object vector cells Spatial memory and navigation require us to learn and remember (Høydal et al., 2019). These spatial signals are all tied to land- the locations of landmarks within our environment. These land- marks, although landmarks in these studies are represented by a marks can take numerous forms from large geographical features range of stimuli from distal room cues to objects close in proxim- to small objects within our local environment. We can use land- ity to the animal. marks to form an allocentric map of the external world that Studies of LEC have shown a clear lack of spatially modu- allows flexible navigation, including the generation of shortcuts; lated signals (Hargreaves et al., 2005; Yoganarasimha et al., the cognitive map (O’Keefe and Nadel, 1978; Tolman, 1948). We 2011), although there is the suggestion that there is weak spatial can also use them to support egocentric representations of the tuning in LEC to local cues within the environment (Neunuebel world that are used during processes such as path integration et al., 2013). This is supported by studies showing that some LEC (McNaughton et al., 2006). In recent years, our understanding of neurons are tuned to objects (Deshmukh et al., 2012; Deshmukh how navigation and spatial memory mechanisms are represented and Knierim, 2011; Tsao et al., 2013). Based on these findings, it in the brain has evolved rapidly, and the circuits supporting both has been suggested that distal global cues could be processed by egocentric and allocentric representations are becoming more well understood. Place cells in the hippocampus fire in consistent locations School of Psychology and Neuroscience, University of St Andrews, St Andrews, UK relative to landmarks providing a potential neural mechanism to Wicking Dementia Research and Education Centre, University of support the cognitive map (O’Keefe and Dostrovsky, 1971). Tasmania, Hobart, TAS, Australia Place cells receive input from two major input pathways from the medial (MEC) and lateral (LEC) entorhinal cortices (Van Strien Corresponding author: et al., 2009). Recent studies of MEC have demonstrated a num- James A. Ainge, School of Psychology and Neuroscience, University of ber of clearly spatially modulated signals. These include grid St Andrews, St Mary’s Quad, St Andrews, Fife, KY16 9JP, UK. cells (Hafting et al., 2005), head direction cells (Sargolini et al., Email: jaa7@st-andrews.ac.uk Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). 2 Brain and Neuroscience Advances MEC, while proximal local cues within the immediate environ- as two different ‘contexts’ by swapping in/out the walls and floor ment are processed by LEC and that these two reference frames of the box. The ‘white’ context had floor and wall inserts painted are tied together in the hippocampus to enable spatial memory white. In the ‘stripes’ context, the walls and floor inserts were and navigation (Knierim et al., 2014; Wang et al., 2020). painted with black and white vertical stripes (5 cm width) with an Consistent with this suggestion, disruption of MEC results in additional plastic-coated metal mesh overlaid on the floor. The deficits in spatial learning and memory when tests use global two objects were attached to the box floor with Dual Lock Velcro cues (Steffenach et al., 2005; Tennant et al., 2018; Van Cauter (3M, St Paul, MN), side-by-side approximately 15 cm apart and et al., 2013) while disruption of LEC impairs learning of a spatial towards the north wall (see Figure 1(a)). Objects used were three- memory task based on local cues (Kuruvilla and Ainge, 2017). dimensional (3D) household items made from ceramic, metal, However, all of the studies covered so far involve testing the glass or plastic that were easy to clean between trials and were ability to use landmarks to support allocentric spatial memory. approximately the size of the rat (in at least one dimension). The Recent studies have demonstrated that LEC neurons show clear box itself was situated on a platform 32 cm above the ground and egocentric coding while MEC signals are dominated by allocen- encircled by a black curtain. Prominent extra-maze cues were tric cues when rats are foraging in an open environment with no attached to the curtain. local cues (Wang et al., 2018). In this study, we asked whether there is a critical role for the LEC in using landmarks to support spatial memory based on either egocentric or allocentric frames Apparatus: experiment 2 of reference. To test the suggestion that LEC is specifically Spatial T-maze task involved in egocentric encoding of space, we examined the effect of LEC lesions on rats’ ability to remember the associations For the spatial task, a modified T-maze with 3 cm high walls was between objects and locations in situations where egocentric and set up 80 cm from the floor. The maze had a central stem that allocentric reference frames were encouraged. For comparison, extended 60 cm to the T-junction before turning off to the left and we examined how these animals performed on a reward-based right arms. Only a portion of the central stem was used for the non-associative spatial task on the T-maze to test the hypothesis experiment (see Figure 1(a)). The maze was open to the testing that LEC may have a general role in spatial processing. Given room, providing animals with prominent extra-maze cues. The that this task can be solved using either an egocentric or allocen- experimenter stood in the same place towards the base of the cen- tric strategy, deficits would suggest an inability to use either type tral stem at the start of every trial to serve as a salient extra-maze of spatial framework. cue. Methods Behavioural procedure: experiment 1 Subjects Following 1 week of extensive handling to habituate the rats to the experimenter, rats were individually habituated to contexts Male Lister Hooded rats (Envigo, Bicester, UK) were housed in (4 days) and then to novel objects within contexts (4 days). The groups of four on a 12-h light/dark cycle (n = 14; average weight two context configurations were not relevant to the current at start of experiment: 359 g). Behavioural testing was conducted hypotheses, and no significant effects of context were found on 5 days a week during the light cycle. The maintenance of labora- task performance (data not reported). The use of the two contexts tory animals and their use in scientific experiments complied and order of presentation of contexts was counterbalanced within with national (Animals (Scientific Procedures) Act, 1986) and and across tasks for both groups of rats. Rats were then tested on international (European Communities Council Directive of 24 a series of recognition tasks in the following order: OR, OP (ego- November 1986 (86/609/EEC)) legislation governing the main- centric) and OP (allocentric). Each task was run for 4 days, with tenance of laboratory animals and their use in scientific experi- rats receiving a sample and test trial on each of those days. At the ments. Local approval was also received from the St Andrews end of the sample trial, rats were placed in a holding cage for Animal Welfare and Ethics Committee. For both experiments, 1 min while the box was cleaned and configured for the test trial. animals had free access to water while in their cages. Specifically for experiment 2, animals were food restricted to no less than 85% of their free feeding bodyweight. This was done to motivate OR task animals during behavioural testing, which involved a food reward. In the sample trial, rats were given 3 min to explore two identical objects (e.g. toy lamps). The trial ended at the end of 3 min or when rats explored both objects for 15 s each, whichever was Apparatus: experiment 1 shorter. During the test trial (3-min duration), rats were exposed to a new copy of the previously seen object (e.g. toy lamp) as Object recognition and object–place well as a novel object not previously seen (e.g. martini glass). associative recognition tasks Memory for the objects from the sample trial would be expected Three recognition memory tasks were conducted: standard object to drive preferential exploration of the novel object (e.g. martini recognition (OR) and two versions of object–place (OP) recogni- glass) in the test trial. Identity and position of the novel object tion promoting an egocentric and allocentric strategy respec- were counterbalanced within and across days for the lesion and tively. All three tasks were run using two objects placed in a sham groups. The experimenter placed the rat in the box facing 67 cm square box with 40 cm high walls. The box could be set up the south wall at the start of both trials (see Figure 1(a)). Kuruvilla et al. 3 (a) Sample Test (d) 2 mins Object recognition (b) Sample Test 2 mins Object-place (egocentric) recognition (c) Sample Test 2 mins Object-place (allocentric) recognition Figure 1. Behavioural task designs. ((a)–(c)) Schematic depicting the structure of a given trial within each behavioural task used in experiment 1. Different novel objects were used each day. Red circles indicate the object/association that is novel. Black arrows indicate the position and direction that the rat was placed in the box at the start of each trial. (d) Schematic showing the design of the modified T-maze used in experiment 2. The area of the maze in grey large dashed lines was blocked off from the rat. The black arrow indicates the position and direction that the rat was placed on the maze at the start of each trial. Circles represent small wells for food rewards. All measurements shown are in cm. and location of novel object were counterbalanced within and OP tasks across days for the lesion and sham groups. In the sample trial, rats were exposed to two different objects (e.g. toy car, clay pot) for 3 min. During the test trial (3-min dura- Behavioural procedure: experiment 2 tion), two identical copies of one of the previously seen objects (e.g. toy cars) were placed in the box for rats to explore. Integrated Spatial T-maze task memory of object and location from the sample trial would be expected to drive preferential exploration of the object in the Rats were habituated to the T-maze for 3 days. On the first day, location in which it had not previously been experienced in the rats were placed at the base of the central stem and allowed to test trial. Two versions of the OP task were run, in which either explore the maze for 5 min. On the second day, a food reward an egocentric or allocentric encoding of space was encouraged. (one-half of a cereal loop, Weeto™) was placed in both the east The key difference between the egocentric and allocentric tasks and west arms of the maze. Rats were placed in the starting point was the direction in which rats were placed at the start of the test and allowed to explore the maze until they had found and eaten trial (see Figure 1(b) and (c)). For the egocentric task, and similar both rewards or 10 min had elapsed, whichever happened sooner. to the OR task, rats were always placed facing the south wall. On the third day, a food reward was only placed in the west arm. However, for the allocentric task, rats began the test trial facing Rats were allowed to find and eat the food reward in the west arm either the east or west walls. This version of the task has previ- and also explore the non-rewarded east arm. ously been used to promote an allocentric encoding of space After completing habituation, rats were trained on a spatial (Langston and Wood, 2010). Initial heading direction and choice task for 7 days, receiving four trials per day. Rats were trained to 35 4 Brain and Neuroscience Advances turn left on to the west arm to receive a food reward (see Figure alcohol (1 min) and finally cleared in xylene. Slides were indi- 1(d)). In every trial, rats were placed at the base of the central vidually removed from xylene and coverslipped using DPX stem and were free to choose to explore east or west arms. If the mountant (BDH Laboratory Supplies, Poole, UK). rat chose the west arm, the trial was ended after the rat had con- sumed the food reward. If the rat chose the east arm, then the rat Lesion analysis was immediately removed from the maze. In between trials, rats were placed in a holding cage while the maze was cleaned and Slides were viewed under a light microscope (Leitz Diaplan) the reward replaced. at magnification ×10 and ×4. The extent of lesioned area was judged by the lack of cell bodies or by cells that were shrunken and damaged. Lesion damage was drawn onto 10 standardised Surgery sections of LEC with reference to Paxinos and Watson (2007; Group sizes were determined based on previous studies show- ranging from −7.66 to −4.42 mm) using Scion Image ing robust effect sizes for rats performing OR and OP tasks in (v4.0.3.2). our laboratory (see also statistics and results sections for details of analysis of generalisability of findings to larger samples). Behavioural analyses Rats in both the lesion (n = 6) and sham (n = 8) groups were ini- tially anaesthetized using isoflurane (Abbot Laboratories Ltd., Discrimination ratio Maidenhead, UK) in an induction box. They were then placed in a stereotaxic frame (David Kopf, Tujunga, CA, USA) where For the three recognition memory tasks, animals were scored to anaesthesia was maintained via a facemask mounted on the be actively exploring an object when their noses were within incisor bar (2%–3% isoflurane, 1.2 l/min O ). A pre-surgical 2 2 cm of the object. The exploration times for the two objects were analgesic Rimadyl (0.05 ml/rat; 5% w/v carprofen; Pfizer Ltd, then converted into discrimination ratios (discrimination Kent, UK) was administered subcutaneously. After shaving the ratio = (time at novel object – time at familiar object) / (time at animal’s scalp, a midline incision was made and holes drilled novel object + time at familiar object)) to determine an animal’s bilaterally at stereotaxic co-ordinates targeting LEC: −6.5 mm relative exploration of the novel versus familiar object or OP from Bregma; ±4.5 mm from the midline (measured on the association. The discrimination ratio calculated here is equiva- skull surface). Dura was cut using the bent tip of a 30-gauge lent to the D2 measure used by Dix and Aggleton (1999). For needle and the pipette lowered into the brain at a 10° angle to each task, discrimination ratios were calculated for each day and 6.4 mm below dura. For animals in the lesion group, 188 nl of then an average across the 4 days used for analysis. To check for ibotenate (0.03M solution in sterile phosphate buffer; Sigma- reliability, a separate observer, who was blind to condition, re- Aldrich, UK) was infused by pressure ejection from a drawn scored a subset of videos for each task, and these scores were glass micropipette (tip diameter 30–40 microns) and left in situ found to be consistently within 10% of the experimenter’s. for 5 min after infusion. Sham controls underwent the identical procedure receiving only the vehicle solution (sterile phosphate buffer). Rats were given 7 days to recover from surgery before Accuracy and latency measures behavioural testing began. For each trial on the spatial T-maze task, animals were judged to have made a choice when all four of their paws were simultane- ously beyond the entrance to either the left or right arms. Animals Perfusion taking a left turn were judged to have made a correct choice while Rats were humanely euthanised with intraperitoneal injections of those turning right were classified as making an incorrect choice. 200 mg/ml/kg sodium pentobarbitone (‘Dolethal’, Univet, Response latencies were also measured on each trial by recording Bicester, UK) and transcardially perfused with phosphate-buff- how long (in seconds) it took animals to make a correct or incor- ered saline (0.9%). This was followed by at least 250 ml of para- rect choice from the time they were placed at the base of the formaldehyde solution (4% made up in 0.1% phosphate buffer central stem. solution). Brains were then extracted and placed overnight in 20% sucrose solution (made up in 0.1% phosphate buffer). Statistical analysis In experiment 1, separate univariate analyses of variance Histology (ANOVAs) were conducted on the discrimination ratios and Brains were immersed in egg yolk within 24-well tissue culture exploration rates in the test phase and sample phase for each of plates containing paraformaldehyde (40%) in the empty neigh- the three recognition memory tasks. To determine the likelihood bouring wells. These were left for 5 days to allow the egg to fix of the reported effects persisting across larger samples, we ran onto the outside of the brains. Brains were subsequently cut into data analysis with bootstrap-coupled estimation (Ho et al., 2019). 50 µm coronal sections on a freezing microtome and then A total of 5000 bootstrap samples were taken; the confidence mounted 1:4 sections onto slides. Sections were then stained on interval is bias-corrected and accelerated. For each permutation P the slides with cresyl violet. To do this, slides were placed in a value, 5000 reshuffles of the control and LEC groups were per- slide holder and then submerged in glass vases of xylene (2 min), formed. The P value reported is the likelihood of observing the 100% alcohol (1 min), 50% alcohol (1 min), water (1 min), cresyl effect size, if the null hypothesis of zero difference is true. Figures violet (2 min), running water (5 min), 50% alcohol (1 min), 100% 3 and 4 along with the statistical analyses presented in Table 1 Kuruvilla et al. 5 were generated from an open-source website (www.estimation- LEC lesions impair egocentric-based stats.com; Ho et al., 2019). One-sample t tests were also used to OP associative recognition assess whether the average discrimination ratios for the lesion As both sham and lesioned rats were able to demonstrate memory and sham groups were different to chance (0) on the various rec- for familiar objects, we could now assess their performance on ognition memory tasks. Additional paired-samples t tests were tasks that required them to remember the association of objects conducted for both groups to compare discrimination ratios and the locations in which they were experienced. Analysis of between the first and second halves of the OP allocentric task. In performance on the OP (egocentric) recognition task also experiment 2, 2 × 7 mixed ANOVAs, with lesion group (LEC; revealed effects that replicated those we have previously reported sham) as the independent factor and training day (days 1–7) as (Wilson et al., 2013b). Average discrimination ratios were sig- the repeated measures factor, were performed for both accuracy nificantly different between groups with LEC lesioned animals and latency during the spatial T-maze task. One-sample t tests performing significantly worse than shams (F = 5.09, p = .023, were conducted to assess whether accuracy performance for the (1,11) η = .316; Figure 3(b)). Rats in the sham group had discrimina- lesion and sham groups was different to chance (0.5) for each test tion ratios significantly greater than chance (t = 3.82, p = .004), day. Bonferroni corrections were applied to post hoc compari- (7) demonstrating that they preferred exploring novel OP associa- sons conducted on significant main and interaction effects. A tions and therefore had remembered familiar OP associations. In Greenhouse–Geisser correction was applied in instances where contrast, rats in the LEC lesion group showed no such preference the sphericity assumption was violated for the repeated measures for exploring novel OP associations (t = –0.81, p = .470), show- factor (training day). Statistical analyses were performed using (4) ing chance-level performance. There was no significant differ- IBM SPSS Statistics 25.0 . ence in the total amount of time exploring objects between rats in sham and LEC lesion groups in either the sample (F = 1.24, (1,11) Results p = .289, η = .101) or test phases (F = 2.37, p = .076, p (1,11) η = .177); Figure 3(e)). Overall, these findings are consistent Histology with previous evidence indicating the role of the LEC in OP asso- ciative recognition memory. Histological analysis determined that five of six rats in the lesion group had successful bilateral lesions of the LEC. Lesion damage from rats with the largest and smallest lesions is depicted in LEC lesions impair allocentric-based Figure 2. One rat had a unilateral lesion of LEC. We have previ- OP associative recognition ously shown that unilateral LEC damage can produce similar deficits to bilateral LEC lesions (Wilson et al., 2013a, 2013b) and Similar to the egocentric version of this task, average discrimina- so analyses were carried out with and without this animal. tion ratios were significantly worse for LEC lesioned animals Analyses presented exclude this animal except where its inclu- relative to shams (F = 12.23, p = .003, η = .526; Figure (1,11) p sion did produce changes to significance levels and these differ- 3(c)). However, unlike in the egocentric version, rats with LEC ences are highlighted. For transparency, the unilateral lesioned lesions were able to recognise OP associations above levels of rats’ data point is specifically labelled in figures. In most rats, chance (t = 2.76, p = .025), as were rats with sham lesions (4) there was some minor damage to ventral subiculum, CA1 and/or (t = 14.43, p < .001). Thus, both groups recognised familiar OP (7) perirhinal cortex but this was estimated at being less than 5% of associations in the test phase. There was no significant difference the overall volume of those structures. Rats in sham group had no in the total amount of time exploring objects between rats in lesion damage. sham and LEC lesion groups in either the sample (F = 0.03, (1,11) p = .870, = .003) or test phases (F = 1.13, p = .156, (1,11) η = .093); Figure 3(f)). Behavioural analysis: experiment 1 To examine the deficit seen in the LEC lesion group in more detail, we compared discrimination ratios between the first and LEC lesions do not impair simple object second halves of the test sessions (Figure 4(a)). One possibility is recognition that rats with LEC lesions initially try and use an egocentric strat- Analysis of performance on the object recognition (OR) task rep- egy, which would be ineffective, and only slowly adapt to using licated our previously reported effects (Kuruvilla and Ainge, an allocentric strategy as the test trial continues. Paired-samples t 2017; Vandrey et al., 2020; Wilson et al., 2013a, 2013b). Average tests revealed that LEC lesioned rats had significantly higher dis- discrimination ratios were not significantly different between crimination ratios in the second half of sessions compared with sham and LEC lesion rats (F = 2.82, p = .061, p = .204; the first (t = –3.47, p = .013) suggesting that these animals were (1,11) (4) Figure 3(a)). In addition, rats in both the sham (t = 8.76, only able to recognise the position of objects within an allocen- (7) p < .001) and lesion (t = 8.93, p < .001) groups had discrimina- tric framework after some time in the environment. This was sup- (4) tion ratios significantly greater than chance. Thus, rats in both ported by one-sample t tests, which revealed that the LEC lesion sham and lesion groups were able to remember objects. There group was performing at chance in the first half of sessions was no significant difference in the total amount of time explor- (t = 1.95, p = .062) but above chance during the second half of (4) ing objects in general (time spent at novel + familiar objects) sessions (t = 3.90, p = .009). Sham-lesioned animals did not (4) between rats in sham and LEC lesion groups (F = 1.19, have significantly different discrimination ratios between the two (1,11) p = .149, p = .098); Figure 3(d)). testing halves (t = 1.71, p = .066). Interestingly, one-sample t (7) CA1 LEC MEC MEC MEC PER PE P R PER PER LEC Po DG DG LEC LEC PER CA1 Po DG DG CA3 PoDG 6 Brain and Neuroscience Advances spent considerably more time exploring objects in first half rela- tive to the second half of test sessions (Figure 4(b)), which would be anticipated as the objects become more familiar. So it is not the case that the general motivational drive to explore objects has been altered by LEC lesions. Rather, it suggests animals’ explora- tory drive decreases in the normal way in both groups, but as the LEC lesioned animals spend more time in the environment, their ability to encode objects in an allocentric framework improves and so despite spending less time exploring the objects, their dis- crimination of novel and familiar OP configurations improves. Overall, rats with LEC lesions remained impaired relative to shams on this OP associative recognition memory task. However, promoting the use of an allocentric spatial framework improved the ability of rats with LEC lesions to recognise an object within a previously experienced location. Furthermore, this ability to recognise an object within a previously experienced location emerges towards the end of the test trial, suggesting that it takes rats with lesions of the LEC longer to place themselves within an allocentric spatial framework. Reproducibility of findings We went on examine the likelihood of the reported effects per- sisting across larger samples. This was done using bootstrap- coupled estimation (Ho et al., 2019). A total of 5000 bootstrap MEC shuffled samples were used to create permutation P values, LEC LEC where P is the likelihood of observing the effect size, if the null hypothesis of zero difference is true. Table 1 illustrates that the findings are robust with the shuffled data confirming the signifi- cant group differences in both OP tasks but not in the OR tasks. Behavioural analysis: experiment 2 LEC lesions do not impair non-associative spatial memory Given that rats were impaired at both types of OP recognition task, we went on to examine whether LEC lesions produced a deficit in a standard test of spatial reference memory on a T-maze. Critically, this task does not require rats to integrate items with spatial locations and rather is a non-associative test of spatial Figure 2. Schematic representation of lesion damage extent. memory. Our previous studies have demonstrated that LEC is Representations of coronal sections adapted from Paxinos and Watson necessary for the association of features of an event suggesting (2007) are at −7.64, −7.04, −6.72, −6.3, −5.8 mm from Bregma, from that non-associative spatial memory should not be affected. LEC top to bottom, respectively. Grey colouring represents lesion damage animals were not impaired in learning this simple spatial task. to LEC where the rat with the greatest lesion damage is shown in light Interestingly, the rats with LEC lesions were more accurate than grey and the rat with the least damage in dark grey. the sham animals at the beginning of training. This was con- firmed with a 2 × 7 mixed ANOVA that revealed significant main effects of group (F = 23.23, p = .001, η = .679) and day (1,11) p tests revealed that sham animals performed above chance in the (F = 3.02, p = .011, η = .216) as well as a significant (6,66) p first half of sessions (t = 6.23, p < .001) but at chance in the lat- (7) group × day (F = 2.39, p = .037, η = 179) interaction. Post (6,66) p ter half (t = 0.64, p = .271). As would be expected, this suggests (7) hoc univariate ANOVAs revealed that LEC lesioned animals that normal animals quickly recognise the position of objects were significantly more accurate than shams on days 1 within an allocentric framework and explore the novel configura- (F = 22.85, p = .001, η = .675) and 2 (F = 5.03, p = .047, (1,11) p (1,11) tion in the early stages of the trial with discrimination becoming η = .314). Post hoc repeated measures ANOVAs revealed a sig- less prominent through the course of the trial as the novel OP nificant difference in accuracy performance across days for the configuration becomes more familiar. These results should also be considered in the context of total sham (F = 6.92, p < .001, η = .497) but not the LEC lesion (6,42) p exploration times for the two groups. Animals in both the sham η group (F = 0.57, p = .638, = .124). One-sample t tests (2.79,11.17) (t = 3.87, p = .003) and LEC lesion groups (t = 10.86, p < .001) further highlighted a difference in accuracy between the two (7) (4) CA1 CA3 CA2 CA1 CA2 DG DG PER CA1 Po DG LEC PER PER LEC LEC PERR PER MEC MEC LEC CA1 Kuruvilla et al. 7 (a) (d) 0.60 80 2.0 2.0 0.50 0.40 1.0 0.0 0.30 0.0 0.20 -2.0 0.10 40 -1.0 0.00 -4.0 30 -2.0 -0.10 -0.20 -3.0 -6.0 -0.30 Control LEC LEC Control LEC LEC N = 8 N = 6 minus N = 8 N = 6 minus Shams Shams Object Recognition Object Recognition (b) (e) 0.50 2.0 1.0 0.40 1.0 60 0.30 0.0 0.20 0.0 0.10 -1.0 -1.0 0.00 -0.10 -2.0 -2.0 -0.20 -3.0 -0.30 Control LEC LEC Control LEC LEC N = 8 N = 6 minus N = 8 N = 6 minus Shams Shams OP Recognition (Egocentric) OP Recognition (Egocentric) 0.60 4.0 0.50 1.0 0.40 2.0 0.30 0.0 0.0 0.20 -2.0 0.10 40 -1.0 0.00 -4.0 -0.10 -2.0 -6.0 -0.20 -8.0 -0.30 Control LEC LEC Control LEC LEC N = 8 N = 6 minus N = 8 N = 6 minus Shams Shams OP Recognition (Allocentric) OP Recognition (Allocentric) Figure 3. Recognition memory of LEC lesion and sham rats in experiment 1. The Cohen’s d difference between LEC and sham is shown in the above Gardner-Altman estimation plot. Both groups are plotted on the left axes; the group mean difference is plotted on floating axes on the right as a bootstrap sampling distribution. The mean difference is depicted as a dot; the 95% confidence interval is indicated by the ends of the vertical error bar. ((a)–(c)) Discrimination ratios across the three recognition memory tasks. ((d)–(f)) Total exploration time across the three recognition memory tasks. One rat with unilateral lesion damage is represented with a black outline. Asterisks represent discrimination ratios significantly different to chance following a one-sample t test (t test vs 0; p < .05). Crosses represent significantly different average discrimination ratios between lesion and sham groups (p < .05). (c) (f) Discrimination ratio Discrimination ratio Discrimination ratio Cohen’s d Cohen’s d Cohen’s d Total exploration time (s) Total exploration time (s) Total exploration time (s) Cohen’s d Cohen’s d Cohen’s d 8 Brain and Neuroscience Advances Control LEC First half of test Second half of test (a) (b) 0.60 50 † † 0.40 0.20 0.00 -0.20 -0.40 -0.60 0 Control LEC Control LEC N = 8 N = 6 N = 8 N = 6 Figure 4. Performance of LEC lesion and sham rats on OP allocentric task split across two halves of testing sessions. (a) Discrimination ratios. (b) Total exploration time. One rat with unilateral lesion damage is represented with a black outline. Asterisks represent discrimination ratios significantly different to chance following a one-sample t test (t test vs 0; p < .05). Crosses represent significantly different average discrimination ratios between lesion and sham groups (p < .05). Table 1. Summary P values and observed effect sizes for differences in discrimination ratios between the two groups across the three recognition memory tasks. Columns 2 and 3 indicate P values and observed effect sizes based on the original data set. Columns 4 and 5 indicate P values and observed effect sizes for 5000 reshuffles of the sham and LEC groups across the three recognition memory tasks. Task Univariate ANOVA Partial η Two-sided permutation Unpaired Cohen’s d between sham t test (n = 8) and lesion (n = 5) groups Object recognition η = .204 P = .132 95.0% CI: −0.96 (−2.79, −0.15) P = .061 OP recognition (egocentric) P = .023 P = .049 95.0% CI: −1.29 (−2.92, 0.01) = .316 OP recognition (allocentric) P = .003 = .526 P = .003 95.0% CI: −1.99 (–3.18, −0.98) OP: object–place; ANOVA: analysis of variance; CI: confidence interval. groups on the first 2 days of training. LEC lesioned animals t = 1.53, p = .170). However, on day 2, the LEC lesion group (7) showed above-chance accuracy (day 1: t = 6.53, p = .003; day 2: was performing above chance (t = 2.91, p = .034) while the (4) (5) t = 3.50, p = .025) while animals in the sham group remained at shams remained at chance levels of accuracy (t = 1.43, p = .197). (4) (7) chance level performance (day 1: t = 1.53, p = .170; day 2: Shams reached above-chance performance by day 3 (t = 7.51, (7) (7) t = 1.43, p = .197). Both groups showed above-chance accuracy p < .001), matching the lesion group (t = 2.74, p = .041). Both (7) (5) on the last 5 days of training. groups showed above-chance accuracy on training days 3 to 7. With the inclusion of the unilateral LEC lesion animal, a 2 × 7 Overall, both sets of analyses indicate that LEC lesioned ani- mixed ANOVA revealed a significant main effect of day mals were unimpaired on a non-associative spatial memory task (F = 4.10, p = .001, η = .255) but neither a main effect of and actually showed improved accuracy relative to shams (6,72) group (F = 4.57, p = .054, = .276) nor a day × group towards the beginning of training. (1,12) (F = 1.78, p = .116, η = .129) interaction effect. Post hoc (6,72) pairwise comparisons on the main effect of ‘Day’ did not reveal a significant difference in accuracy between training days for LEC lesions do not impair spatial either the sham or lesion groups. One-sample t tests highlighted a response latencies difference in accuracy between the two groups during the initial days of training (see Figure 5(a)). Both groups showed at-chance Response latencies were compared between the two groups accuracy on training day 1 (LEC: t = 2.45, p = .058, shams: across training days as a proxy for evaluating decision-making (5) Discrimination ratio Total exploration time (s) Kuruvilla et al. 9 (a) (b) * * * * 60 * Control (N = 8) LEC (N = 6) 0 0 1 23 45 6 7 1 2 3 45 6 7 Day Day Figure 5. Performance of LEC lesion and sham rats during a non-associative spatial task on a modified T-maze. (a) Average correct responses across 7 training days. (b) Average trial completion time from the rat being placed on the maze to touching reward across 7 training days (four trials per day). Asterisks represent accuracy performance significantly above chance following a one-sample t test (t test vs 0.5; p < .05). certainty and motivation towards the food reward. Figure 5(b) Results demonstrated that LEC is critical for remembering the shows that both groups decreased their response latency across location of objects within an environment irrespective of the testing. A 2 × 7 mixed ANOVA revealed a significant main effect frame of reference in which they are presented. of day (F = 3.23, p = .039, η = .227) but neither a main On first inspection, this argues against the suggestion that (2.81,30.94) effect of group (F = 0.16, p = .701, η = .014) nor a LEC is preferentially involved in processing egocentric rather (1,11) day × group (F = 1.24, p = .311, η = .101) interaction than allocentric spatial information and rather suggests that the (2.81,30.94) (Figure 5(b)). Post hoc pairwise comparisons revealed that both principle role of LEC is the integration of features of an event. groups were faster at accurately completing trials on day 7 com- However, closer inspection of the data reveals that while LEC pared to day 1 (M = –8.77, SE = 1.98, p = .021). These results indi- lesioned rats are impaired at the allocentric version of the OP cate that both the LEC lesion animals and shams were similarly task, relative to shams, they are still performing above chance. motivated and decisive in completing the spatial task. In addi- Interestingly, LEC lesioned animals’ ability to recognise a famil- tion, the change in latency between the first and last training day iar OP association within an allocentric spatial framework took for both groups suggests that animals were learning the task over longer to develop than controls with significant memory for OP time, as expected. The significance of the data did not change associations only occurring in the second half of the allocentric when the animal with the unilateral lesion was included. OP test trials. This shows that they can still use an allocentric frame of reference to remember the association of object with location even if it is not as efficient a process as it is in controls. Discussion In contrast, performance of the LEC lesioned rats on the egocen- The hippocampal–entorhinal network has been shown to be criti- tric version of the task is at chance. Clearly, while LEC is critical cal for spatial memory (Ainge et al., 2006, 2007; Ainge and for the association of object and place, it is particularly important Langston, 2012; Andersen et al., 2006; Broadbent et al., 2004; for tasks that involve processing of spatial information in an ego- Martin and Clark, 2007; Morris et al., 1982; Save and Sargolini, centric frame of reference. 2017; Steffenach et al., 2005; Van Cauter et al., 2013). Recent How is it possible for rats with LEC lesions to remember the studies, however, have suggested that different parts of the net- allocentric location of a previously experienced object? Previous work are tuned to either egocentric or allocentric frames of refer- experiments using complete LEC lesions had shown deficits in ence with LEC showing clear egocentric spatial tuning (Wang all associative recognition memory tasks (Chao et al., 2016; et al., 2018). We have previously shown that LEC is critically Kuruvilla and Ainge, 2017; Rodo et al., 2017; Van Cauter et al., important for integrating features of an event, including the loca- 2013; Wilson et al., 2013a, 2013b). However, recent studies tions of objects within an environment (Kuruvilla and Ainge, using more specific manipulations have demonstrated the exist- 2017; Wilson et al., 2013a, 2013b). Here, we tested the role of ence of segregated functional circuits within the LEC. Leitner LEC in integrating objects within either an egocentric or allocen- et al. (2016) showed that reelin and calbindin positive cells in tric frame of reference. This was assessed using two versions of layer 2 of LEC respond differently to odours and Vandrey et al. the OP recognition task: one in which animals were introduced to (2020) went on to show that specific inactivation of reelin posi- the environment from a consistent spatial location to encourage tive cells in layer 2a of LEC results in impaired object–place– an egocentric frame of reference, the other in which animals are context memory while leaving object–context memory intact. introduced from multiple different locations encouraging an allo- While neither of these studies tested different spatial reference centric frame of reference. Rats would have to first situate them- frames, they do demonstrate that LEC as a whole does not act as selves in more global, allocentric space before being able to a functional unit but rather that it is made up of specialised sub- make an associative memory judgement as the positions of the systems that can be functionally segregated. One potential expla- objects would have moved relative to an egocentric framework. nation for the spared ability of LEC lesioned rats in this study to Accuracy (%) Trial completion time (s) 10 Brain and Neuroscience Advances remember allocentric OP associations is that all of the lesioned implications for the current study. First, it would be consistent rats had some residual LEC tissue within a consistent functional with the primary role for LEC being in integration of features of unit that could be used to support allocentric representations of an event and memory for the associations between these features. the location of objects. Histological analysis of the lesions in this Second, close inspection of the anatomical inputs to LEC show a study showed that LEC lesioned rats did consistently have por- strong input from olfactory areas. Olfactory stimuli will be much tions of the most ventromedial part of LEC still intact. However, more salient for local rather than global cues which would again while it is possible that ventromedial LEC has a specific role in be consistent with LEC having a role in the processing of local allocentric spatial processing, this has not been investigated. spatial features. Future studies could examine whether there are differences in Overall, the current findings are in agreement with previous processing egocentric versus allocentric frames of reference studies suggesting that the central role of LEC is in the integra- across this ventromedial–dorsolateral band of LEC. tion of features that make up episodic memory (Beer et al., 2013; We went onto examine whether lesions of LEC would pro- Chao et al., 2016; Hunsaker et al., 2013; Rodo et al., 2017; Van duce a general deficit in spatial memory by examining perfor- Cauter et al., 2013; Wilson et al., 2013a, 2013b). When the find- mance on a non-associative reference memory task on the ings are combined with anatomical and electrophysiological T-maze. The data show that rats with lesions of the LEC were not studies, it creates a clear picture in which LEC encodes local, impaired and were actually better than controls in the early stages multimodal stimuli such as objects within the environment. This of the task. It is possible for rats to use either egocentric or allo- information can be used to support either egocentric or allocen- centric frames of reference to solve the T-maze task and so it is tric frames of reference. interesting to examine whether the associative recognition tasks give an insight into how LEC lesions affected rats’ ability to Acknowledgements remember the correct spatial location on the T-maze. Results We would like to thank Nayantara Kansal and Judita Huber in collecting from the OP experiments show that LEC animals do not remem- behavioural data. ber the egocentric position of a previously seen object but do remember its allocentric position, albeit less well than shams. Declaration of conflicting interests One interesting possibility here is that normal animals spend the The author(s) declared no potential conflicts of interest with respect to first few days of T-maze training learning both an egocentric and the research, authorship, and/or publication of this article. allocentric frame of reference while LEC lesioned rats rely solely on the allocentric. This would mean they have less information to Funding learn, making the task easier which could explain the increased T-maze accuracy in the LEC lesioned rats. Future studies could The author(s) disclosed receipt of the following financial support for the incorporate rotation trials where allocentric and egocentric research, authorship, and/or publication of this article: This work was frames of reference are placed into conflict to examine whether supported by Biotechnology and Biological Sciences Research Council (BBSRC) grant BB/I019367/1. disruption of LEC changes strategy use. Previous studies have suggested that LEC may process local spatial information rather than distal, global cues. We have previ- ORCID iDs ously shown that LEC lesioned rats can remember the position of Maneesh V. Kuruvilla https://orcid.org/0000-0002-3045-1529 food rewards in relation to global allocentric space but are James A. Ainge https://orcid.org/0000-0002-0007-1533 impaired when learning the position of food relative to local cues (Kuruvilla and Ainge, 2017). The current data are consistent with References this suggestion. In the allocentric OP task, animals are introduced Ainge JA and Langston RF (2012) Ontogeny of neural circuits underly- to the environment from multiple different locations. This will ing spatial memory in the rat. Frontiers in Neural Circuits 6: 8. mean that they encode the position of the objects within the allo- Ainge JA, Heron-Maxwell C, Theofilas P, et al. (2006) The role of the centric framework of the testing arena but also the global spatial hippocampus in object recognition in rats: Examination of the influ- cues from the testing room. The fact that the LEC lesioned rats do ence of task parameters and lesion size. Behavioural Brain Research better in the allocentric condition may be due to their ability to 167(1): 183–195. use the global distal room cues to solve the task. Ainge JA, Van der Meer MAA, Langston RF, et al. (2007) Exploring the The current data are consistent with unit recording data that role of context-dependent hippocampal activity in spatial alternation have shown LEC responses to objects within the environment behavior. Hippocampus 17(10): 988–1002. (Deshmukh et al., 2012; Deshmukh and Knierim, 2011; Tsao Andersen P, Morris R, Amaral D, et al. (2006) The Hippocampus Book. et al., 2013; Keene et al., 2016). These studies show that LEC Oxford: Oxford University Press. Barry C, Lever C, Hayman R, et al. (2006) The boundary vector cell neurons develop specific and consistent spatial tuning when model of place cell firing and spatial memory. Reviews in the Neuro- objects are included in the local environment. Tsao et al. (2013) sciences 17(1–2): 71–97. went on to show a small subset of LEC neurons encode the posi- Beer Z, Chwiesko C, Kitsukawa T, et al. (2013) Spatial and stimulus- tions in which objects have been previously experienced. These type tuning in the LEC, MEC, POR, PrC, CA1, and CA3 during ‘trace’ cells are a cellular correlate of OP memory and could sup- spontaneous item recognition memory. Hippocampus 23(12): port the behaviour reported in this study. 1425–1438. Previous studies have shown that LEC is one of four cortical Bota M, Sporns O and Swanson LW (2015) Architecture of the cerebral hubs which receives extensive connections from the rest of cortical association connectome underlying cognition. Proceedings cortex (Bota et al., 2015) putting it in an ideal position to inte- of the National Academy of Sciences of the United States of America grate signals from multiple areas of the brain. This has two 112(16): E2093–E2101. Kuruvilla et al. 11 Broadbent NJ, Squire LR and Clark RE (2004) Spatial memory, recog- Neunuebel JP, Yoganarasimha D, Rao G, et al. (2013) Conflicts between nition memory, and the hippocampus. Proceedings of the National local and global spatial frameworks dissociate neural representations Academy of Sciences 101(40): 14515–14520. of the lateral and medial entorhinal cortex. The Journal of Neurosci- Chao OY, Huston JP, Li JS, et al. (2016) The medial prefrontal cor- ence: The Official Journal of the Society for Neuroscience 33(22): tex-lateral entorhinal cortex circuit is essential for episodic-like 9246–9258. memory and associative object-recognition. Hippocampus 26(5): O’Keefe J and Dostrovsky J (1971) The hippocampus as a spatial map: 633–645. Preliminary evidence from unit activity in the freely-moving rat. Deshmukh SS and Knierim JJ (2011) Representation of non-spatial and Brain Research 34(1): 171–175. spatial information in the lateral entorhinal cortex. Frontiers in O’Keefe J and Nadel L (1978) The Hippocampus as a Cognitive Map. Behavioral Neuroscience 5: 69. Oxford: Clarendon Press. Deshmukh SS, Johnson JL and Knierim JJ (2012) Perirhinal cortex rep- Paxinos G and Watson C (2007) The Rat Brain in Stereotaxic Coordi- resents nonspatial, but not spatial, information in rats foraging in the nates. Amsterdam: Academic/Elsevier. presence of objects: Comparison with lateral entorhinal cortex. Hip- Rodo C, Sargolini F and Save E (2017) Processing of spatial and non- pocampus 22(10): 2045–2058. spatial information in rats with lesions of the medial and lateral ento- Dix SL and Aggleton JP (1999) Extending the spontaneous preference rhinal cortex: Environmental complexity matters. Behavioural Brain test of recognition: Evidence of object-location and object-context Research 320: 200–209. recognition. Behavioural Brain Research 99(2): 191–200. Sargolini F, Fyhn M, Hafting T, et al. (2006) Conjunctive representa- Hafting T, Fyhn M, Molden S, et al. (2005) Microstructure of a spatial tion of position, direction, and velocity in entorhinal cortex. Science map in the entorhinal cortex. Nature 436(7052): 801–806. 312(5774): 758–762. Hargreaves EL, Rao G, Lee I, et al. (2005) Major dissociation between Save E and Sargolini F (2017) Disentangling the role of the MEC and LEC medial and lateral entorhinal input to dorsal hippocampus. Science in the processing of spatial and non-spatial information: Contribution 308(5729): 1792–1794. of lesion studies. Frontiers in Systems Neuroscience. Available at: Ho J, Tumkaya T, Aryal S, et al. (2019) Moving beyond P values: Data https://www.frontiersin.org/article/10.3389/fnsys.2017.00081 analysis with estimation graphics. Nature Methods 16(7): 565–566. Solstad T, Boccara CN, Kropff E, et al. (2008) Representation of geometric Høydal OA, Skytøen ER, Andersson SO, et al. (2019) Object-vector cod- borders in the entorhinal cortex. Science 322(5909): 1865–1868. ing in the medial entorhinal cortex. Nature 568(7752): 400–404. Steffenach H-A, Witter M, Moser M-B, et al. (2005) Spatial memory in Hunsaker MR, Chen V, Tran GT, et al. (2013) The medial and lateral the rat requires the dorsolateral band of the entorhinal cortex. Neuron entorhinal cortex both contribute to contextual and item recognition 45(2): 301–313. memory: A test of the binding of items and context model. Hippo- Tennant SA, Fischer L, Garden DLF, et al. (2018) Stellate cells in the campus 23(5): 380–391. medial entorhinal cortex are required for spatial learning. Cell Keene CS, Bladon J, McKenzie S, et al. (2016) Complementary func- Reports 22(5): 1313–1324. tional organization of neuronal activity patterns in the perirhinal, Tolman EC (1948) Cognitive maps in rats and man. Psychological lateral entorhinal, and medial entorhinal cortices. The Journal of Review 55(4): 189–208. Neuroscience: The Official Journal of the Society for Neuroscience Tsao A, Moser MB and Moser EI (2013) Traces of experience in the 36(13): 3660–3675. lateral entorhinal cortex. Current Biology 23(5): 399–405. Knierim JJ, Neunuebel JP and Deshmukh SS (2014) Functional correlates Van Cauter T, Camon J, Alvernhe A, et al. (2013) Distinct roles of medial of the lateral and medial entorhinal cortex: Objects, path integration and lateral entorhinal cortex in spatial cognition. Cerebral Cortex and local-global reference frames. Philosophical Transactions of the 23(2): 451–459. Royal Society 369(1635): 20130369. Van Strien NM, Cappaert NLM and Witter MP (2009) The anatomy of Kuruvilla MV and Ainge JA (2017) Lateral entorhinal cortex lesions memory: An interactive overview of the parahippocampal-hippo- impair local spatial frameworks. Frontiers in Systems Neurosci- campal network. Nature Reviews Neuroscience 10(4): 272–282. ence. Available at: https://www.frontiersin.org/article/10.3389/ Vandrey B, Garden DLF, Ambrozova V, et al. (2020) Fan cells in layer 2 fnsys.2017.00030 of the lateral entorhinal cortex are critical for episodic-like memory. Langston RF and Wood ER (2010) Associative recognition and the hippo- Current Biology 30(1): 169–175. campus: Differential effects of hippocampal lesions on object-place, Wang C, Chen X and Knierim JJ (2020) Egocentric and allocentric rep- object-context and object-place-context memory. Hippocampus resentations of space in the rodent brain. Current Opinion in Neuro- 20(10): 1139–1153. biology 60: 12–20. Leitner FC, Melzer S, Lütcke H, et al. (2016) Spatially segregated feed- Wang C, Chen X, Lee H, et al. (2018) Egocentric coding of external forward and feedback neurons support differential odor processing in items in the lateral entorhinal cortex. Science 362(6417): 945–949. the lateral entorhinal cortex. Nature Neuroscience 19(7): 935–944. Wilson DIG, Langston RF, Schlesiger MI, et al. (2013a) Lateral entorhi- McNaughton BL, Battaglia FP, Jensen O, et al. (2006) Path integration nal cortex is critical for novel object-context recognition. Hippocam- and the neural basis of the ‘cognitive map’. Nature Reviews Neuro- pus 23(5): 352–366. science 7(8): 663–678. Wilson DIG, Watanabe S, Milner H, et al. (2013b) Lateral entorhinal Martin SJ and Clark RE (2007) The rodent hippocampus and spatial cortex is necessary for associative but not nonassociative recognition memory: From synapses to systems. Cellular and Molecular Life memory. Hippocampus 23(12): 1280–1290. Sciences 64(4): 401. Yoganarasimha D, Rao G and Knierim JJ (2011) Lateral entorhinal neu- Morris RGM, Garrud P, Rawlins JNP, et al. (1982) Place navigation rons are not spatially selective in cue-rich environments. Hippocam- impaired in rats with hippocampal lesions. Nature 297(5868): 681–683. pus 21(12): 1363–1374. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Brain and Neuroscience Advances SAGE

Lateral entorhinal cortex lesions impair both egocentric and allocentric object–place associations:

Loading next page...
 
/lp/sage/lateral-entorhinal-cortex-lesions-impair-both-egocentric-and-77hSYeUtzH
Publisher
SAGE
Copyright
Copyright © 2022 by SAGE Publications Ltd and British Neuroscience Association, unless otherwise noted. Manuscript content on this site is licensed under Creative Commons Licenses
ISSN
2398-2128
eISSN
2398-2128
DOI
10.1177/2398212820939463
Publisher site
See Article on Publisher Site

Abstract

During navigation, landmark processing is critical either for generating an allocentric-based cognitive map or in facilitating egocentric-based strategies. Increasing evidence from manipulation and single-unit recording studies has highlighted the role of the entorhinal cortex in processing landmarks. In particular, the lateral (LEC) and medial (MEC) sub-regions of the entorhinal cortex have been shown to attend to proximal and distal landmarks, respectively. Recent studies have identified a further dissociation in cue processing between the LEC and MEC based on spatial frames of reference. Neurons in the LEC preferentially encode egocentric cues while those in the MEC encode allocentric cues. In this study, we assessed the impact of disrupting the LEC on landmark-based spatial memory in both egocentric and allocentric reference frames. Animals that received excitotoxic lesions of the LEC were significantly impaired, relative to controls, on both egocentric and allocentric versions of an object–place association task. Notably, LEC lesioned animals performed at chance on the egocentric version but above chance on the allocentric version. There was no significant difference in performance between the two groups on an object recognition and spatial T-maze task. Taken together, these results indicate that the LEC plays a role in feature integration more broadly and in specifically processing spatial information within an egocentric reference frame. Keywords Hippocampus, spatial memory, associative, episodic memory, navigation, medial entorhinal cortex, cognitive map, landmarks Received: 1 April 2020; accepted: 11 June 2020 2006), border cells (Barry et al., 2006; Solstad et al., 2008), con- Introduction junctive cells (Sargolini et al., 2006), and object vector cells Spatial memory and navigation require us to learn and remember (Høydal et al., 2019). These spatial signals are all tied to land- the locations of landmarks within our environment. These land- marks, although landmarks in these studies are represented by a marks can take numerous forms from large geographical features range of stimuli from distal room cues to objects close in proxim- to small objects within our local environment. We can use land- ity to the animal. marks to form an allocentric map of the external world that Studies of LEC have shown a clear lack of spatially modu- allows flexible navigation, including the generation of shortcuts; lated signals (Hargreaves et al., 2005; Yoganarasimha et al., the cognitive map (O’Keefe and Nadel, 1978; Tolman, 1948). We 2011), although there is the suggestion that there is weak spatial can also use them to support egocentric representations of the tuning in LEC to local cues within the environment (Neunuebel world that are used during processes such as path integration et al., 2013). This is supported by studies showing that some LEC (McNaughton et al., 2006). In recent years, our understanding of neurons are tuned to objects (Deshmukh et al., 2012; Deshmukh how navigation and spatial memory mechanisms are represented and Knierim, 2011; Tsao et al., 2013). Based on these findings, it in the brain has evolved rapidly, and the circuits supporting both has been suggested that distal global cues could be processed by egocentric and allocentric representations are becoming more well understood. Place cells in the hippocampus fire in consistent locations School of Psychology and Neuroscience, University of St Andrews, St Andrews, UK relative to landmarks providing a potential neural mechanism to Wicking Dementia Research and Education Centre, University of support the cognitive map (O’Keefe and Dostrovsky, 1971). Tasmania, Hobart, TAS, Australia Place cells receive input from two major input pathways from the medial (MEC) and lateral (LEC) entorhinal cortices (Van Strien Corresponding author: et al., 2009). Recent studies of MEC have demonstrated a num- James A. Ainge, School of Psychology and Neuroscience, University of ber of clearly spatially modulated signals. These include grid St Andrews, St Mary’s Quad, St Andrews, Fife, KY16 9JP, UK. cells (Hafting et al., 2005), head direction cells (Sargolini et al., Email: jaa7@st-andrews.ac.uk Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). 2 Brain and Neuroscience Advances MEC, while proximal local cues within the immediate environ- as two different ‘contexts’ by swapping in/out the walls and floor ment are processed by LEC and that these two reference frames of the box. The ‘white’ context had floor and wall inserts painted are tied together in the hippocampus to enable spatial memory white. In the ‘stripes’ context, the walls and floor inserts were and navigation (Knierim et al., 2014; Wang et al., 2020). painted with black and white vertical stripes (5 cm width) with an Consistent with this suggestion, disruption of MEC results in additional plastic-coated metal mesh overlaid on the floor. The deficits in spatial learning and memory when tests use global two objects were attached to the box floor with Dual Lock Velcro cues (Steffenach et al., 2005; Tennant et al., 2018; Van Cauter (3M, St Paul, MN), side-by-side approximately 15 cm apart and et al., 2013) while disruption of LEC impairs learning of a spatial towards the north wall (see Figure 1(a)). Objects used were three- memory task based on local cues (Kuruvilla and Ainge, 2017). dimensional (3D) household items made from ceramic, metal, However, all of the studies covered so far involve testing the glass or plastic that were easy to clean between trials and were ability to use landmarks to support allocentric spatial memory. approximately the size of the rat (in at least one dimension). The Recent studies have demonstrated that LEC neurons show clear box itself was situated on a platform 32 cm above the ground and egocentric coding while MEC signals are dominated by allocen- encircled by a black curtain. Prominent extra-maze cues were tric cues when rats are foraging in an open environment with no attached to the curtain. local cues (Wang et al., 2018). In this study, we asked whether there is a critical role for the LEC in using landmarks to support spatial memory based on either egocentric or allocentric frames Apparatus: experiment 2 of reference. To test the suggestion that LEC is specifically Spatial T-maze task involved in egocentric encoding of space, we examined the effect of LEC lesions on rats’ ability to remember the associations For the spatial task, a modified T-maze with 3 cm high walls was between objects and locations in situations where egocentric and set up 80 cm from the floor. The maze had a central stem that allocentric reference frames were encouraged. For comparison, extended 60 cm to the T-junction before turning off to the left and we examined how these animals performed on a reward-based right arms. Only a portion of the central stem was used for the non-associative spatial task on the T-maze to test the hypothesis experiment (see Figure 1(a)). The maze was open to the testing that LEC may have a general role in spatial processing. Given room, providing animals with prominent extra-maze cues. The that this task can be solved using either an egocentric or allocen- experimenter stood in the same place towards the base of the cen- tric strategy, deficits would suggest an inability to use either type tral stem at the start of every trial to serve as a salient extra-maze of spatial framework. cue. Methods Behavioural procedure: experiment 1 Subjects Following 1 week of extensive handling to habituate the rats to the experimenter, rats were individually habituated to contexts Male Lister Hooded rats (Envigo, Bicester, UK) were housed in (4 days) and then to novel objects within contexts (4 days). The groups of four on a 12-h light/dark cycle (n = 14; average weight two context configurations were not relevant to the current at start of experiment: 359 g). Behavioural testing was conducted hypotheses, and no significant effects of context were found on 5 days a week during the light cycle. The maintenance of labora- task performance (data not reported). The use of the two contexts tory animals and their use in scientific experiments complied and order of presentation of contexts was counterbalanced within with national (Animals (Scientific Procedures) Act, 1986) and and across tasks for both groups of rats. Rats were then tested on international (European Communities Council Directive of 24 a series of recognition tasks in the following order: OR, OP (ego- November 1986 (86/609/EEC)) legislation governing the main- centric) and OP (allocentric). Each task was run for 4 days, with tenance of laboratory animals and their use in scientific experi- rats receiving a sample and test trial on each of those days. At the ments. Local approval was also received from the St Andrews end of the sample trial, rats were placed in a holding cage for Animal Welfare and Ethics Committee. For both experiments, 1 min while the box was cleaned and configured for the test trial. animals had free access to water while in their cages. Specifically for experiment 2, animals were food restricted to no less than 85% of their free feeding bodyweight. This was done to motivate OR task animals during behavioural testing, which involved a food reward. In the sample trial, rats were given 3 min to explore two identical objects (e.g. toy lamps). The trial ended at the end of 3 min or when rats explored both objects for 15 s each, whichever was Apparatus: experiment 1 shorter. During the test trial (3-min duration), rats were exposed to a new copy of the previously seen object (e.g. toy lamp) as Object recognition and object–place well as a novel object not previously seen (e.g. martini glass). associative recognition tasks Memory for the objects from the sample trial would be expected Three recognition memory tasks were conducted: standard object to drive preferential exploration of the novel object (e.g. martini recognition (OR) and two versions of object–place (OP) recogni- glass) in the test trial. Identity and position of the novel object tion promoting an egocentric and allocentric strategy respec- were counterbalanced within and across days for the lesion and tively. All three tasks were run using two objects placed in a sham groups. The experimenter placed the rat in the box facing 67 cm square box with 40 cm high walls. The box could be set up the south wall at the start of both trials (see Figure 1(a)). Kuruvilla et al. 3 (a) Sample Test (d) 2 mins Object recognition (b) Sample Test 2 mins Object-place (egocentric) recognition (c) Sample Test 2 mins Object-place (allocentric) recognition Figure 1. Behavioural task designs. ((a)–(c)) Schematic depicting the structure of a given trial within each behavioural task used in experiment 1. Different novel objects were used each day. Red circles indicate the object/association that is novel. Black arrows indicate the position and direction that the rat was placed in the box at the start of each trial. (d) Schematic showing the design of the modified T-maze used in experiment 2. The area of the maze in grey large dashed lines was blocked off from the rat. The black arrow indicates the position and direction that the rat was placed on the maze at the start of each trial. Circles represent small wells for food rewards. All measurements shown are in cm. and location of novel object were counterbalanced within and OP tasks across days for the lesion and sham groups. In the sample trial, rats were exposed to two different objects (e.g. toy car, clay pot) for 3 min. During the test trial (3-min dura- Behavioural procedure: experiment 2 tion), two identical copies of one of the previously seen objects (e.g. toy cars) were placed in the box for rats to explore. Integrated Spatial T-maze task memory of object and location from the sample trial would be expected to drive preferential exploration of the object in the Rats were habituated to the T-maze for 3 days. On the first day, location in which it had not previously been experienced in the rats were placed at the base of the central stem and allowed to test trial. Two versions of the OP task were run, in which either explore the maze for 5 min. On the second day, a food reward an egocentric or allocentric encoding of space was encouraged. (one-half of a cereal loop, Weeto™) was placed in both the east The key difference between the egocentric and allocentric tasks and west arms of the maze. Rats were placed in the starting point was the direction in which rats were placed at the start of the test and allowed to explore the maze until they had found and eaten trial (see Figure 1(b) and (c)). For the egocentric task, and similar both rewards or 10 min had elapsed, whichever happened sooner. to the OR task, rats were always placed facing the south wall. On the third day, a food reward was only placed in the west arm. However, for the allocentric task, rats began the test trial facing Rats were allowed to find and eat the food reward in the west arm either the east or west walls. This version of the task has previ- and also explore the non-rewarded east arm. ously been used to promote an allocentric encoding of space After completing habituation, rats were trained on a spatial (Langston and Wood, 2010). Initial heading direction and choice task for 7 days, receiving four trials per day. Rats were trained to 35 4 Brain and Neuroscience Advances turn left on to the west arm to receive a food reward (see Figure alcohol (1 min) and finally cleared in xylene. Slides were indi- 1(d)). In every trial, rats were placed at the base of the central vidually removed from xylene and coverslipped using DPX stem and were free to choose to explore east or west arms. If the mountant (BDH Laboratory Supplies, Poole, UK). rat chose the west arm, the trial was ended after the rat had con- sumed the food reward. If the rat chose the east arm, then the rat Lesion analysis was immediately removed from the maze. In between trials, rats were placed in a holding cage while the maze was cleaned and Slides were viewed under a light microscope (Leitz Diaplan) the reward replaced. at magnification ×10 and ×4. The extent of lesioned area was judged by the lack of cell bodies or by cells that were shrunken and damaged. Lesion damage was drawn onto 10 standardised Surgery sections of LEC with reference to Paxinos and Watson (2007; Group sizes were determined based on previous studies show- ranging from −7.66 to −4.42 mm) using Scion Image ing robust effect sizes for rats performing OR and OP tasks in (v4.0.3.2). our laboratory (see also statistics and results sections for details of analysis of generalisability of findings to larger samples). Behavioural analyses Rats in both the lesion (n = 6) and sham (n = 8) groups were ini- tially anaesthetized using isoflurane (Abbot Laboratories Ltd., Discrimination ratio Maidenhead, UK) in an induction box. They were then placed in a stereotaxic frame (David Kopf, Tujunga, CA, USA) where For the three recognition memory tasks, animals were scored to anaesthesia was maintained via a facemask mounted on the be actively exploring an object when their noses were within incisor bar (2%–3% isoflurane, 1.2 l/min O ). A pre-surgical 2 2 cm of the object. The exploration times for the two objects were analgesic Rimadyl (0.05 ml/rat; 5% w/v carprofen; Pfizer Ltd, then converted into discrimination ratios (discrimination Kent, UK) was administered subcutaneously. After shaving the ratio = (time at novel object – time at familiar object) / (time at animal’s scalp, a midline incision was made and holes drilled novel object + time at familiar object)) to determine an animal’s bilaterally at stereotaxic co-ordinates targeting LEC: −6.5 mm relative exploration of the novel versus familiar object or OP from Bregma; ±4.5 mm from the midline (measured on the association. The discrimination ratio calculated here is equiva- skull surface). Dura was cut using the bent tip of a 30-gauge lent to the D2 measure used by Dix and Aggleton (1999). For needle and the pipette lowered into the brain at a 10° angle to each task, discrimination ratios were calculated for each day and 6.4 mm below dura. For animals in the lesion group, 188 nl of then an average across the 4 days used for analysis. To check for ibotenate (0.03M solution in sterile phosphate buffer; Sigma- reliability, a separate observer, who was blind to condition, re- Aldrich, UK) was infused by pressure ejection from a drawn scored a subset of videos for each task, and these scores were glass micropipette (tip diameter 30–40 microns) and left in situ found to be consistently within 10% of the experimenter’s. for 5 min after infusion. Sham controls underwent the identical procedure receiving only the vehicle solution (sterile phosphate buffer). Rats were given 7 days to recover from surgery before Accuracy and latency measures behavioural testing began. For each trial on the spatial T-maze task, animals were judged to have made a choice when all four of their paws were simultane- ously beyond the entrance to either the left or right arms. Animals Perfusion taking a left turn were judged to have made a correct choice while Rats were humanely euthanised with intraperitoneal injections of those turning right were classified as making an incorrect choice. 200 mg/ml/kg sodium pentobarbitone (‘Dolethal’, Univet, Response latencies were also measured on each trial by recording Bicester, UK) and transcardially perfused with phosphate-buff- how long (in seconds) it took animals to make a correct or incor- ered saline (0.9%). This was followed by at least 250 ml of para- rect choice from the time they were placed at the base of the formaldehyde solution (4% made up in 0.1% phosphate buffer central stem. solution). Brains were then extracted and placed overnight in 20% sucrose solution (made up in 0.1% phosphate buffer). Statistical analysis In experiment 1, separate univariate analyses of variance Histology (ANOVAs) were conducted on the discrimination ratios and Brains were immersed in egg yolk within 24-well tissue culture exploration rates in the test phase and sample phase for each of plates containing paraformaldehyde (40%) in the empty neigh- the three recognition memory tasks. To determine the likelihood bouring wells. These were left for 5 days to allow the egg to fix of the reported effects persisting across larger samples, we ran onto the outside of the brains. Brains were subsequently cut into data analysis with bootstrap-coupled estimation (Ho et al., 2019). 50 µm coronal sections on a freezing microtome and then A total of 5000 bootstrap samples were taken; the confidence mounted 1:4 sections onto slides. Sections were then stained on interval is bias-corrected and accelerated. For each permutation P the slides with cresyl violet. To do this, slides were placed in a value, 5000 reshuffles of the control and LEC groups were per- slide holder and then submerged in glass vases of xylene (2 min), formed. The P value reported is the likelihood of observing the 100% alcohol (1 min), 50% alcohol (1 min), water (1 min), cresyl effect size, if the null hypothesis of zero difference is true. Figures violet (2 min), running water (5 min), 50% alcohol (1 min), 100% 3 and 4 along with the statistical analyses presented in Table 1 Kuruvilla et al. 5 were generated from an open-source website (www.estimation- LEC lesions impair egocentric-based stats.com; Ho et al., 2019). One-sample t tests were also used to OP associative recognition assess whether the average discrimination ratios for the lesion As both sham and lesioned rats were able to demonstrate memory and sham groups were different to chance (0) on the various rec- for familiar objects, we could now assess their performance on ognition memory tasks. Additional paired-samples t tests were tasks that required them to remember the association of objects conducted for both groups to compare discrimination ratios and the locations in which they were experienced. Analysis of between the first and second halves of the OP allocentric task. In performance on the OP (egocentric) recognition task also experiment 2, 2 × 7 mixed ANOVAs, with lesion group (LEC; revealed effects that replicated those we have previously reported sham) as the independent factor and training day (days 1–7) as (Wilson et al., 2013b). Average discrimination ratios were sig- the repeated measures factor, were performed for both accuracy nificantly different between groups with LEC lesioned animals and latency during the spatial T-maze task. One-sample t tests performing significantly worse than shams (F = 5.09, p = .023, were conducted to assess whether accuracy performance for the (1,11) η = .316; Figure 3(b)). Rats in the sham group had discrimina- lesion and sham groups was different to chance (0.5) for each test tion ratios significantly greater than chance (t = 3.82, p = .004), day. Bonferroni corrections were applied to post hoc compari- (7) demonstrating that they preferred exploring novel OP associa- sons conducted on significant main and interaction effects. A tions and therefore had remembered familiar OP associations. In Greenhouse–Geisser correction was applied in instances where contrast, rats in the LEC lesion group showed no such preference the sphericity assumption was violated for the repeated measures for exploring novel OP associations (t = –0.81, p = .470), show- factor (training day). Statistical analyses were performed using (4) ing chance-level performance. There was no significant differ- IBM SPSS Statistics 25.0 . ence in the total amount of time exploring objects between rats in sham and LEC lesion groups in either the sample (F = 1.24, (1,11) Results p = .289, η = .101) or test phases (F = 2.37, p = .076, p (1,11) η = .177); Figure 3(e)). Overall, these findings are consistent Histology with previous evidence indicating the role of the LEC in OP asso- ciative recognition memory. Histological analysis determined that five of six rats in the lesion group had successful bilateral lesions of the LEC. Lesion damage from rats with the largest and smallest lesions is depicted in LEC lesions impair allocentric-based Figure 2. One rat had a unilateral lesion of LEC. We have previ- OP associative recognition ously shown that unilateral LEC damage can produce similar deficits to bilateral LEC lesions (Wilson et al., 2013a, 2013b) and Similar to the egocentric version of this task, average discrimina- so analyses were carried out with and without this animal. tion ratios were significantly worse for LEC lesioned animals Analyses presented exclude this animal except where its inclu- relative to shams (F = 12.23, p = .003, η = .526; Figure (1,11) p sion did produce changes to significance levels and these differ- 3(c)). However, unlike in the egocentric version, rats with LEC ences are highlighted. For transparency, the unilateral lesioned lesions were able to recognise OP associations above levels of rats’ data point is specifically labelled in figures. In most rats, chance (t = 2.76, p = .025), as were rats with sham lesions (4) there was some minor damage to ventral subiculum, CA1 and/or (t = 14.43, p < .001). Thus, both groups recognised familiar OP (7) perirhinal cortex but this was estimated at being less than 5% of associations in the test phase. There was no significant difference the overall volume of those structures. Rats in sham group had no in the total amount of time exploring objects between rats in lesion damage. sham and LEC lesion groups in either the sample (F = 0.03, (1,11) p = .870, = .003) or test phases (F = 1.13, p = .156, (1,11) η = .093); Figure 3(f)). Behavioural analysis: experiment 1 To examine the deficit seen in the LEC lesion group in more detail, we compared discrimination ratios between the first and LEC lesions do not impair simple object second halves of the test sessions (Figure 4(a)). One possibility is recognition that rats with LEC lesions initially try and use an egocentric strat- Analysis of performance on the object recognition (OR) task rep- egy, which would be ineffective, and only slowly adapt to using licated our previously reported effects (Kuruvilla and Ainge, an allocentric strategy as the test trial continues. Paired-samples t 2017; Vandrey et al., 2020; Wilson et al., 2013a, 2013b). Average tests revealed that LEC lesioned rats had significantly higher dis- discrimination ratios were not significantly different between crimination ratios in the second half of sessions compared with sham and LEC lesion rats (F = 2.82, p = .061, p = .204; the first (t = –3.47, p = .013) suggesting that these animals were (1,11) (4) Figure 3(a)). In addition, rats in both the sham (t = 8.76, only able to recognise the position of objects within an allocen- (7) p < .001) and lesion (t = 8.93, p < .001) groups had discrimina- tric framework after some time in the environment. This was sup- (4) tion ratios significantly greater than chance. Thus, rats in both ported by one-sample t tests, which revealed that the LEC lesion sham and lesion groups were able to remember objects. There group was performing at chance in the first half of sessions was no significant difference in the total amount of time explor- (t = 1.95, p = .062) but above chance during the second half of (4) ing objects in general (time spent at novel + familiar objects) sessions (t = 3.90, p = .009). Sham-lesioned animals did not (4) between rats in sham and LEC lesion groups (F = 1.19, have significantly different discrimination ratios between the two (1,11) p = .149, p = .098); Figure 3(d)). testing halves (t = 1.71, p = .066). Interestingly, one-sample t (7) CA1 LEC MEC MEC MEC PER PE P R PER PER LEC Po DG DG LEC LEC PER CA1 Po DG DG CA3 PoDG 6 Brain and Neuroscience Advances spent considerably more time exploring objects in first half rela- tive to the second half of test sessions (Figure 4(b)), which would be anticipated as the objects become more familiar. So it is not the case that the general motivational drive to explore objects has been altered by LEC lesions. Rather, it suggests animals’ explora- tory drive decreases in the normal way in both groups, but as the LEC lesioned animals spend more time in the environment, their ability to encode objects in an allocentric framework improves and so despite spending less time exploring the objects, their dis- crimination of novel and familiar OP configurations improves. Overall, rats with LEC lesions remained impaired relative to shams on this OP associative recognition memory task. However, promoting the use of an allocentric spatial framework improved the ability of rats with LEC lesions to recognise an object within a previously experienced location. Furthermore, this ability to recognise an object within a previously experienced location emerges towards the end of the test trial, suggesting that it takes rats with lesions of the LEC longer to place themselves within an allocentric spatial framework. Reproducibility of findings We went on examine the likelihood of the reported effects per- sisting across larger samples. This was done using bootstrap- coupled estimation (Ho et al., 2019). A total of 5000 bootstrap MEC shuffled samples were used to create permutation P values, LEC LEC where P is the likelihood of observing the effect size, if the null hypothesis of zero difference is true. Table 1 illustrates that the findings are robust with the shuffled data confirming the signifi- cant group differences in both OP tasks but not in the OR tasks. Behavioural analysis: experiment 2 LEC lesions do not impair non-associative spatial memory Given that rats were impaired at both types of OP recognition task, we went on to examine whether LEC lesions produced a deficit in a standard test of spatial reference memory on a T-maze. Critically, this task does not require rats to integrate items with spatial locations and rather is a non-associative test of spatial Figure 2. Schematic representation of lesion damage extent. memory. Our previous studies have demonstrated that LEC is Representations of coronal sections adapted from Paxinos and Watson necessary for the association of features of an event suggesting (2007) are at −7.64, −7.04, −6.72, −6.3, −5.8 mm from Bregma, from that non-associative spatial memory should not be affected. LEC top to bottom, respectively. Grey colouring represents lesion damage animals were not impaired in learning this simple spatial task. to LEC where the rat with the greatest lesion damage is shown in light Interestingly, the rats with LEC lesions were more accurate than grey and the rat with the least damage in dark grey. the sham animals at the beginning of training. This was con- firmed with a 2 × 7 mixed ANOVA that revealed significant main effects of group (F = 23.23, p = .001, η = .679) and day (1,11) p tests revealed that sham animals performed above chance in the (F = 3.02, p = .011, η = .216) as well as a significant (6,66) p first half of sessions (t = 6.23, p < .001) but at chance in the lat- (7) group × day (F = 2.39, p = .037, η = 179) interaction. Post (6,66) p ter half (t = 0.64, p = .271). As would be expected, this suggests (7) hoc univariate ANOVAs revealed that LEC lesioned animals that normal animals quickly recognise the position of objects were significantly more accurate than shams on days 1 within an allocentric framework and explore the novel configura- (F = 22.85, p = .001, η = .675) and 2 (F = 5.03, p = .047, (1,11) p (1,11) tion in the early stages of the trial with discrimination becoming η = .314). Post hoc repeated measures ANOVAs revealed a sig- less prominent through the course of the trial as the novel OP nificant difference in accuracy performance across days for the configuration becomes more familiar. These results should also be considered in the context of total sham (F = 6.92, p < .001, η = .497) but not the LEC lesion (6,42) p exploration times for the two groups. Animals in both the sham η group (F = 0.57, p = .638, = .124). One-sample t tests (2.79,11.17) (t = 3.87, p = .003) and LEC lesion groups (t = 10.86, p < .001) further highlighted a difference in accuracy between the two (7) (4) CA1 CA3 CA2 CA1 CA2 DG DG PER CA1 Po DG LEC PER PER LEC LEC PERR PER MEC MEC LEC CA1 Kuruvilla et al. 7 (a) (d) 0.60 80 2.0 2.0 0.50 0.40 1.0 0.0 0.30 0.0 0.20 -2.0 0.10 40 -1.0 0.00 -4.0 30 -2.0 -0.10 -0.20 -3.0 -6.0 -0.30 Control LEC LEC Control LEC LEC N = 8 N = 6 minus N = 8 N = 6 minus Shams Shams Object Recognition Object Recognition (b) (e) 0.50 2.0 1.0 0.40 1.0 60 0.30 0.0 0.20 0.0 0.10 -1.0 -1.0 0.00 -0.10 -2.0 -2.0 -0.20 -3.0 -0.30 Control LEC LEC Control LEC LEC N = 8 N = 6 minus N = 8 N = 6 minus Shams Shams OP Recognition (Egocentric) OP Recognition (Egocentric) 0.60 4.0 0.50 1.0 0.40 2.0 0.30 0.0 0.0 0.20 -2.0 0.10 40 -1.0 0.00 -4.0 -0.10 -2.0 -6.0 -0.20 -8.0 -0.30 Control LEC LEC Control LEC LEC N = 8 N = 6 minus N = 8 N = 6 minus Shams Shams OP Recognition (Allocentric) OP Recognition (Allocentric) Figure 3. Recognition memory of LEC lesion and sham rats in experiment 1. The Cohen’s d difference between LEC and sham is shown in the above Gardner-Altman estimation plot. Both groups are plotted on the left axes; the group mean difference is plotted on floating axes on the right as a bootstrap sampling distribution. The mean difference is depicted as a dot; the 95% confidence interval is indicated by the ends of the vertical error bar. ((a)–(c)) Discrimination ratios across the three recognition memory tasks. ((d)–(f)) Total exploration time across the three recognition memory tasks. One rat with unilateral lesion damage is represented with a black outline. Asterisks represent discrimination ratios significantly different to chance following a one-sample t test (t test vs 0; p < .05). Crosses represent significantly different average discrimination ratios between lesion and sham groups (p < .05). (c) (f) Discrimination ratio Discrimination ratio Discrimination ratio Cohen’s d Cohen’s d Cohen’s d Total exploration time (s) Total exploration time (s) Total exploration time (s) Cohen’s d Cohen’s d Cohen’s d 8 Brain and Neuroscience Advances Control LEC First half of test Second half of test (a) (b) 0.60 50 † † 0.40 0.20 0.00 -0.20 -0.40 -0.60 0 Control LEC Control LEC N = 8 N = 6 N = 8 N = 6 Figure 4. Performance of LEC lesion and sham rats on OP allocentric task split across two halves of testing sessions. (a) Discrimination ratios. (b) Total exploration time. One rat with unilateral lesion damage is represented with a black outline. Asterisks represent discrimination ratios significantly different to chance following a one-sample t test (t test vs 0; p < .05). Crosses represent significantly different average discrimination ratios between lesion and sham groups (p < .05). Table 1. Summary P values and observed effect sizes for differences in discrimination ratios between the two groups across the three recognition memory tasks. Columns 2 and 3 indicate P values and observed effect sizes based on the original data set. Columns 4 and 5 indicate P values and observed effect sizes for 5000 reshuffles of the sham and LEC groups across the three recognition memory tasks. Task Univariate ANOVA Partial η Two-sided permutation Unpaired Cohen’s d between sham t test (n = 8) and lesion (n = 5) groups Object recognition η = .204 P = .132 95.0% CI: −0.96 (−2.79, −0.15) P = .061 OP recognition (egocentric) P = .023 P = .049 95.0% CI: −1.29 (−2.92, 0.01) = .316 OP recognition (allocentric) P = .003 = .526 P = .003 95.0% CI: −1.99 (–3.18, −0.98) OP: object–place; ANOVA: analysis of variance; CI: confidence interval. groups on the first 2 days of training. LEC lesioned animals t = 1.53, p = .170). However, on day 2, the LEC lesion group (7) showed above-chance accuracy (day 1: t = 6.53, p = .003; day 2: was performing above chance (t = 2.91, p = .034) while the (4) (5) t = 3.50, p = .025) while animals in the sham group remained at shams remained at chance levels of accuracy (t = 1.43, p = .197). (4) (7) chance level performance (day 1: t = 1.53, p = .170; day 2: Shams reached above-chance performance by day 3 (t = 7.51, (7) (7) t = 1.43, p = .197). Both groups showed above-chance accuracy p < .001), matching the lesion group (t = 2.74, p = .041). Both (7) (5) on the last 5 days of training. groups showed above-chance accuracy on training days 3 to 7. With the inclusion of the unilateral LEC lesion animal, a 2 × 7 Overall, both sets of analyses indicate that LEC lesioned ani- mixed ANOVA revealed a significant main effect of day mals were unimpaired on a non-associative spatial memory task (F = 4.10, p = .001, η = .255) but neither a main effect of and actually showed improved accuracy relative to shams (6,72) group (F = 4.57, p = .054, = .276) nor a day × group towards the beginning of training. (1,12) (F = 1.78, p = .116, η = .129) interaction effect. Post hoc (6,72) pairwise comparisons on the main effect of ‘Day’ did not reveal a significant difference in accuracy between training days for LEC lesions do not impair spatial either the sham or lesion groups. One-sample t tests highlighted a response latencies difference in accuracy between the two groups during the initial days of training (see Figure 5(a)). Both groups showed at-chance Response latencies were compared between the two groups accuracy on training day 1 (LEC: t = 2.45, p = .058, shams: across training days as a proxy for evaluating decision-making (5) Discrimination ratio Total exploration time (s) Kuruvilla et al. 9 (a) (b) * * * * 60 * Control (N = 8) LEC (N = 6) 0 0 1 23 45 6 7 1 2 3 45 6 7 Day Day Figure 5. Performance of LEC lesion and sham rats during a non-associative spatial task on a modified T-maze. (a) Average correct responses across 7 training days. (b) Average trial completion time from the rat being placed on the maze to touching reward across 7 training days (four trials per day). Asterisks represent accuracy performance significantly above chance following a one-sample t test (t test vs 0.5; p < .05). certainty and motivation towards the food reward. Figure 5(b) Results demonstrated that LEC is critical for remembering the shows that both groups decreased their response latency across location of objects within an environment irrespective of the testing. A 2 × 7 mixed ANOVA revealed a significant main effect frame of reference in which they are presented. of day (F = 3.23, p = .039, η = .227) but neither a main On first inspection, this argues against the suggestion that (2.81,30.94) effect of group (F = 0.16, p = .701, η = .014) nor a LEC is preferentially involved in processing egocentric rather (1,11) day × group (F = 1.24, p = .311, η = .101) interaction than allocentric spatial information and rather suggests that the (2.81,30.94) (Figure 5(b)). Post hoc pairwise comparisons revealed that both principle role of LEC is the integration of features of an event. groups were faster at accurately completing trials on day 7 com- However, closer inspection of the data reveals that while LEC pared to day 1 (M = –8.77, SE = 1.98, p = .021). These results indi- lesioned rats are impaired at the allocentric version of the OP cate that both the LEC lesion animals and shams were similarly task, relative to shams, they are still performing above chance. motivated and decisive in completing the spatial task. In addi- Interestingly, LEC lesioned animals’ ability to recognise a famil- tion, the change in latency between the first and last training day iar OP association within an allocentric spatial framework took for both groups suggests that animals were learning the task over longer to develop than controls with significant memory for OP time, as expected. The significance of the data did not change associations only occurring in the second half of the allocentric when the animal with the unilateral lesion was included. OP test trials. This shows that they can still use an allocentric frame of reference to remember the association of object with location even if it is not as efficient a process as it is in controls. Discussion In contrast, performance of the LEC lesioned rats on the egocen- The hippocampal–entorhinal network has been shown to be criti- tric version of the task is at chance. Clearly, while LEC is critical cal for spatial memory (Ainge et al., 2006, 2007; Ainge and for the association of object and place, it is particularly important Langston, 2012; Andersen et al., 2006; Broadbent et al., 2004; for tasks that involve processing of spatial information in an ego- Martin and Clark, 2007; Morris et al., 1982; Save and Sargolini, centric frame of reference. 2017; Steffenach et al., 2005; Van Cauter et al., 2013). Recent How is it possible for rats with LEC lesions to remember the studies, however, have suggested that different parts of the net- allocentric location of a previously experienced object? Previous work are tuned to either egocentric or allocentric frames of refer- experiments using complete LEC lesions had shown deficits in ence with LEC showing clear egocentric spatial tuning (Wang all associative recognition memory tasks (Chao et al., 2016; et al., 2018). We have previously shown that LEC is critically Kuruvilla and Ainge, 2017; Rodo et al., 2017; Van Cauter et al., important for integrating features of an event, including the loca- 2013; Wilson et al., 2013a, 2013b). However, recent studies tions of objects within an environment (Kuruvilla and Ainge, using more specific manipulations have demonstrated the exist- 2017; Wilson et al., 2013a, 2013b). Here, we tested the role of ence of segregated functional circuits within the LEC. Leitner LEC in integrating objects within either an egocentric or allocen- et al. (2016) showed that reelin and calbindin positive cells in tric frame of reference. This was assessed using two versions of layer 2 of LEC respond differently to odours and Vandrey et al. the OP recognition task: one in which animals were introduced to (2020) went on to show that specific inactivation of reelin posi- the environment from a consistent spatial location to encourage tive cells in layer 2a of LEC results in impaired object–place– an egocentric frame of reference, the other in which animals are context memory while leaving object–context memory intact. introduced from multiple different locations encouraging an allo- While neither of these studies tested different spatial reference centric frame of reference. Rats would have to first situate them- frames, they do demonstrate that LEC as a whole does not act as selves in more global, allocentric space before being able to a functional unit but rather that it is made up of specialised sub- make an associative memory judgement as the positions of the systems that can be functionally segregated. One potential expla- objects would have moved relative to an egocentric framework. nation for the spared ability of LEC lesioned rats in this study to Accuracy (%) Trial completion time (s) 10 Brain and Neuroscience Advances remember allocentric OP associations is that all of the lesioned implications for the current study. First, it would be consistent rats had some residual LEC tissue within a consistent functional with the primary role for LEC being in integration of features of unit that could be used to support allocentric representations of an event and memory for the associations between these features. the location of objects. Histological analysis of the lesions in this Second, close inspection of the anatomical inputs to LEC show a study showed that LEC lesioned rats did consistently have por- strong input from olfactory areas. Olfactory stimuli will be much tions of the most ventromedial part of LEC still intact. However, more salient for local rather than global cues which would again while it is possible that ventromedial LEC has a specific role in be consistent with LEC having a role in the processing of local allocentric spatial processing, this has not been investigated. spatial features. Future studies could examine whether there are differences in Overall, the current findings are in agreement with previous processing egocentric versus allocentric frames of reference studies suggesting that the central role of LEC is in the integra- across this ventromedial–dorsolateral band of LEC. tion of features that make up episodic memory (Beer et al., 2013; We went onto examine whether lesions of LEC would pro- Chao et al., 2016; Hunsaker et al., 2013; Rodo et al., 2017; Van duce a general deficit in spatial memory by examining perfor- Cauter et al., 2013; Wilson et al., 2013a, 2013b). When the find- mance on a non-associative reference memory task on the ings are combined with anatomical and electrophysiological T-maze. The data show that rats with lesions of the LEC were not studies, it creates a clear picture in which LEC encodes local, impaired and were actually better than controls in the early stages multimodal stimuli such as objects within the environment. This of the task. It is possible for rats to use either egocentric or allo- information can be used to support either egocentric or allocen- centric frames of reference to solve the T-maze task and so it is tric frames of reference. interesting to examine whether the associative recognition tasks give an insight into how LEC lesions affected rats’ ability to Acknowledgements remember the correct spatial location on the T-maze. Results We would like to thank Nayantara Kansal and Judita Huber in collecting from the OP experiments show that LEC animals do not remem- behavioural data. ber the egocentric position of a previously seen object but do remember its allocentric position, albeit less well than shams. Declaration of conflicting interests One interesting possibility here is that normal animals spend the The author(s) declared no potential conflicts of interest with respect to first few days of T-maze training learning both an egocentric and the research, authorship, and/or publication of this article. allocentric frame of reference while LEC lesioned rats rely solely on the allocentric. This would mean they have less information to Funding learn, making the task easier which could explain the increased T-maze accuracy in the LEC lesioned rats. Future studies could The author(s) disclosed receipt of the following financial support for the incorporate rotation trials where allocentric and egocentric research, authorship, and/or publication of this article: This work was frames of reference are placed into conflict to examine whether supported by Biotechnology and Biological Sciences Research Council (BBSRC) grant BB/I019367/1. disruption of LEC changes strategy use. Previous studies have suggested that LEC may process local spatial information rather than distal, global cues. We have previ- ORCID iDs ously shown that LEC lesioned rats can remember the position of Maneesh V. Kuruvilla https://orcid.org/0000-0002-3045-1529 food rewards in relation to global allocentric space but are James A. Ainge https://orcid.org/0000-0002-0007-1533 impaired when learning the position of food relative to local cues (Kuruvilla and Ainge, 2017). The current data are consistent with References this suggestion. In the allocentric OP task, animals are introduced Ainge JA and Langston RF (2012) Ontogeny of neural circuits underly- to the environment from multiple different locations. This will ing spatial memory in the rat. Frontiers in Neural Circuits 6: 8. mean that they encode the position of the objects within the allo- Ainge JA, Heron-Maxwell C, Theofilas P, et al. (2006) The role of the centric framework of the testing arena but also the global spatial hippocampus in object recognition in rats: Examination of the influ- cues from the testing room. The fact that the LEC lesioned rats do ence of task parameters and lesion size. Behavioural Brain Research better in the allocentric condition may be due to their ability to 167(1): 183–195. use the global distal room cues to solve the task. Ainge JA, Van der Meer MAA, Langston RF, et al. (2007) Exploring the The current data are consistent with unit recording data that role of context-dependent hippocampal activity in spatial alternation have shown LEC responses to objects within the environment behavior. Hippocampus 17(10): 988–1002. (Deshmukh et al., 2012; Deshmukh and Knierim, 2011; Tsao Andersen P, Morris R, Amaral D, et al. (2006) The Hippocampus Book. et al., 2013; Keene et al., 2016). These studies show that LEC Oxford: Oxford University Press. Barry C, Lever C, Hayman R, et al. (2006) The boundary vector cell neurons develop specific and consistent spatial tuning when model of place cell firing and spatial memory. Reviews in the Neuro- objects are included in the local environment. Tsao et al. (2013) sciences 17(1–2): 71–97. went on to show a small subset of LEC neurons encode the posi- Beer Z, Chwiesko C, Kitsukawa T, et al. (2013) Spatial and stimulus- tions in which objects have been previously experienced. These type tuning in the LEC, MEC, POR, PrC, CA1, and CA3 during ‘trace’ cells are a cellular correlate of OP memory and could sup- spontaneous item recognition memory. Hippocampus 23(12): port the behaviour reported in this study. 1425–1438. Previous studies have shown that LEC is one of four cortical Bota M, Sporns O and Swanson LW (2015) Architecture of the cerebral hubs which receives extensive connections from the rest of cortical association connectome underlying cognition. Proceedings cortex (Bota et al., 2015) putting it in an ideal position to inte- of the National Academy of Sciences of the United States of America grate signals from multiple areas of the brain. This has two 112(16): E2093–E2101. Kuruvilla et al. 11 Broadbent NJ, Squire LR and Clark RE (2004) Spatial memory, recog- Neunuebel JP, Yoganarasimha D, Rao G, et al. (2013) Conflicts between nition memory, and the hippocampus. Proceedings of the National local and global spatial frameworks dissociate neural representations Academy of Sciences 101(40): 14515–14520. of the lateral and medial entorhinal cortex. The Journal of Neurosci- Chao OY, Huston JP, Li JS, et al. (2016) The medial prefrontal cor- ence: The Official Journal of the Society for Neuroscience 33(22): tex-lateral entorhinal cortex circuit is essential for episodic-like 9246–9258. memory and associative object-recognition. Hippocampus 26(5): O’Keefe J and Dostrovsky J (1971) The hippocampus as a spatial map: 633–645. Preliminary evidence from unit activity in the freely-moving rat. Deshmukh SS and Knierim JJ (2011) Representation of non-spatial and Brain Research 34(1): 171–175. spatial information in the lateral entorhinal cortex. Frontiers in O’Keefe J and Nadel L (1978) The Hippocampus as a Cognitive Map. Behavioral Neuroscience 5: 69. Oxford: Clarendon Press. Deshmukh SS, Johnson JL and Knierim JJ (2012) Perirhinal cortex rep- Paxinos G and Watson C (2007) The Rat Brain in Stereotaxic Coordi- resents nonspatial, but not spatial, information in rats foraging in the nates. Amsterdam: Academic/Elsevier. presence of objects: Comparison with lateral entorhinal cortex. Hip- Rodo C, Sargolini F and Save E (2017) Processing of spatial and non- pocampus 22(10): 2045–2058. spatial information in rats with lesions of the medial and lateral ento- Dix SL and Aggleton JP (1999) Extending the spontaneous preference rhinal cortex: Environmental complexity matters. Behavioural Brain test of recognition: Evidence of object-location and object-context Research 320: 200–209. recognition. Behavioural Brain Research 99(2): 191–200. Sargolini F, Fyhn M, Hafting T, et al. (2006) Conjunctive representa- Hafting T, Fyhn M, Molden S, et al. (2005) Microstructure of a spatial tion of position, direction, and velocity in entorhinal cortex. Science map in the entorhinal cortex. Nature 436(7052): 801–806. 312(5774): 758–762. Hargreaves EL, Rao G, Lee I, et al. (2005) Major dissociation between Save E and Sargolini F (2017) Disentangling the role of the MEC and LEC medial and lateral entorhinal input to dorsal hippocampus. Science in the processing of spatial and non-spatial information: Contribution 308(5729): 1792–1794. of lesion studies. Frontiers in Systems Neuroscience. Available at: Ho J, Tumkaya T, Aryal S, et al. (2019) Moving beyond P values: Data https://www.frontiersin.org/article/10.3389/fnsys.2017.00081 analysis with estimation graphics. Nature Methods 16(7): 565–566. Solstad T, Boccara CN, Kropff E, et al. (2008) Representation of geometric Høydal OA, Skytøen ER, Andersson SO, et al. (2019) Object-vector cod- borders in the entorhinal cortex. Science 322(5909): 1865–1868. ing in the medial entorhinal cortex. Nature 568(7752): 400–404. Steffenach H-A, Witter M, Moser M-B, et al. (2005) Spatial memory in Hunsaker MR, Chen V, Tran GT, et al. (2013) The medial and lateral the rat requires the dorsolateral band of the entorhinal cortex. Neuron entorhinal cortex both contribute to contextual and item recognition 45(2): 301–313. memory: A test of the binding of items and context model. Hippo- Tennant SA, Fischer L, Garden DLF, et al. (2018) Stellate cells in the campus 23(5): 380–391. medial entorhinal cortex are required for spatial learning. Cell Keene CS, Bladon J, McKenzie S, et al. (2016) Complementary func- Reports 22(5): 1313–1324. tional organization of neuronal activity patterns in the perirhinal, Tolman EC (1948) Cognitive maps in rats and man. Psychological lateral entorhinal, and medial entorhinal cortices. The Journal of Review 55(4): 189–208. Neuroscience: The Official Journal of the Society for Neuroscience Tsao A, Moser MB and Moser EI (2013) Traces of experience in the 36(13): 3660–3675. lateral entorhinal cortex. Current Biology 23(5): 399–405. Knierim JJ, Neunuebel JP and Deshmukh SS (2014) Functional correlates Van Cauter T, Camon J, Alvernhe A, et al. (2013) Distinct roles of medial of the lateral and medial entorhinal cortex: Objects, path integration and lateral entorhinal cortex in spatial cognition. Cerebral Cortex and local-global reference frames. Philosophical Transactions of the 23(2): 451–459. Royal Society 369(1635): 20130369. Van Strien NM, Cappaert NLM and Witter MP (2009) The anatomy of Kuruvilla MV and Ainge JA (2017) Lateral entorhinal cortex lesions memory: An interactive overview of the parahippocampal-hippo- impair local spatial frameworks. Frontiers in Systems Neurosci- campal network. Nature Reviews Neuroscience 10(4): 272–282. ence. Available at: https://www.frontiersin.org/article/10.3389/ Vandrey B, Garden DLF, Ambrozova V, et al. (2020) Fan cells in layer 2 fnsys.2017.00030 of the lateral entorhinal cortex are critical for episodic-like memory. Langston RF and Wood ER (2010) Associative recognition and the hippo- Current Biology 30(1): 169–175. campus: Differential effects of hippocampal lesions on object-place, Wang C, Chen X and Knierim JJ (2020) Egocentric and allocentric rep- object-context and object-place-context memory. Hippocampus resentations of space in the rodent brain. Current Opinion in Neuro- 20(10): 1139–1153. biology 60: 12–20. Leitner FC, Melzer S, Lütcke H, et al. (2016) Spatially segregated feed- Wang C, Chen X, Lee H, et al. (2018) Egocentric coding of external forward and feedback neurons support differential odor processing in items in the lateral entorhinal cortex. Science 362(6417): 945–949. the lateral entorhinal cortex. Nature Neuroscience 19(7): 935–944. Wilson DIG, Langston RF, Schlesiger MI, et al. (2013a) Lateral entorhi- McNaughton BL, Battaglia FP, Jensen O, et al. (2006) Path integration nal cortex is critical for novel object-context recognition. Hippocam- and the neural basis of the ‘cognitive map’. Nature Reviews Neuro- pus 23(5): 352–366. science 7(8): 663–678. Wilson DIG, Watanabe S, Milner H, et al. (2013b) Lateral entorhinal Martin SJ and Clark RE (2007) The rodent hippocampus and spatial cortex is necessary for associative but not nonassociative recognition memory: From synapses to systems. Cellular and Molecular Life memory. Hippocampus 23(12): 1280–1290. Sciences 64(4): 401. Yoganarasimha D, Rao G and Knierim JJ (2011) Lateral entorhinal neu- Morris RGM, Garrud P, Rawlins JNP, et al. (1982) Place navigation rons are not spatially selective in cue-rich environments. Hippocam- impaired in rats with hippocampal lesions. Nature 297(5868): 681–683. pus 21(12): 1363–1374.

Journal

Brain and Neuroscience AdvancesSAGE

Published: Jul 14, 2020

Keywords: Hippocampus; spatial memory; associative; episodic memory; navigation; medial entorhinal cortex; cognitive map; landmarks

References