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

Learn More →

Does EEG Montage Influence Alzheimer's Disease Electroclinic Diagnosis?

Does EEG Montage Influence Alzheimer's Disease Electroclinic Diagnosis? SAGE-Hindawi Access to Research International Journal of Alzheimer’s Disease Volume 2011, Article ID 761891, 6 pages doi:10.4061/2011/761891 Research Article Does EEG Montage Influence Alzheimer’s Disease Electroclinic Diagnosis? 1 1 2 3 3 3 L. R. Trambaiolli, A. C. Lorena, F. J. Fraga, P. A. M. K. Kanda, R. Nitrini, and R. Anghinah Mathematics, Computing and Cognition Center (CMCC), Universidade Federal do ABC (UFABC), Rua Santa Adelia, 166, 09210-170 Santo Andre, SP, Brazil Engineering, Modeling and Applied Social Sciences Center (CECS), Universidade Federal do ABC (UFABC), Rua Santa Adelia, 166, 09210-170 Santo Andre, SP, Brazil Reference Center of Behavioral Disturbances and Dementia (CEREDIC) and Neurology, Department of Medicine School of University of Sao ˜ Paulo (USP), Rua Arruda Alvim, 206, 05.410-020 Sao ˜ Paulo, SP, Brazil Correspondence should be addressed to F. J. Fraga, franciscojfraga@gmail.com Received 23 December 2010; Revised 23 February 2011; Accepted 7 March 2011 Academic Editor: Fabrizio Vecchio Copyright © 2011 L. R. Trambaiolli et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. There is not a specific Alzheimer’s disease (AD) diagnostic test. AD diagnosis relies on clinical history, neuropsychological, and laboratory tests, neuroimaging and electroencephalography. Therefore, new approaches are necessary to enable earlier and more accurate diagnosis and to measure treatment results. Quantitative EEG (qEEG) can be used as a diagnostic tool in selected cases. The aim of this study was to answer if distinct electrode montages have different sensitivity when differentiating controls from AD patients. We analyzed EEG spectral peaks (delta, theta, alpha, beta, and gamma bands), and we compared references (Biauricular, Longitudinal bipolar, Crossed bipolar, Counterpart bipolar, and Cz reference). Support Vector Machines and Logistic Regression classifiers showed Counterpart bipolar montage as the most sensitive electrode combination. Our results suggest that Counterpart bipolar montage is the best choice to study EEG spectral peaks of controls versus AD. 1. Introduction [10–13] in AD EEG background. Saletu et al. [14]found a localized temporal decrease of alpha and beta activities in Alzheimer’s disease (AD) diagnosis is based upon clinical AD and slow cerebral rhythms widespread distribution in history, neuropsychological and laboratory tests, neuroimag- vascular dementia (VaD) [10–13]. Pucci et al. [15] proposed that a decrease in alpha frequency to 6.0–8.0 Hz could be an ing, and electroencephalography (EEG). New approaches are necessary to earlier and more accurate diagnosis [1, 2]and to AD marker. measure treatment results [3]. Despite the knowledge grounded in this field during the last decades, there are lots of unanswered questions that hin- EEG visual analysis can be a helpful diagnostic test in AD [4, 5]. Background frequency displacement to delta and theta der qEEG consolidation as an AD diagnostic tool. Our objec- frequencies and the dropout of central alpha rhythm are tive was to study if distinct electrode montages have different sensitivity when differentiating controls from AD patients. common EEG findings in AD [6]. Accordingly, Sandmann et al. [7] observed a direct correlation between the degree of cognitive impairment and the power of low-frequency 2. Materials and Methods electrical activity in the EEG. Since the first quantitative EEG (qEEG) studies by 2.1. Subjects. The dataset was composed of electroencep- Lehmann [8]and Duffy et al. [9], spectral analysis (specA) halograms (EEGs) recorded from two groups aged from 60 and statistics have been applied to EEG. Moreover, specA to 80 years: (S1) 12 normal subjects and (S2) 22 probable has been considered from 71% to 81% sensitive to changes AD patients (NINCDS-ADRDA criteria) [16]. AD group 2 International Journal of Alzheimer’s Disease was classified as having mild to moderate symptoms (DSM- 4.0 to 8.0 Hz), alpha (from 8.0 to 12.0 Hz), beta (from 12.0 IV-TR) [17]. Both groups were submitted to the Brazilian to 30.0 Hz), and gamma (from 30.0 to 50.0 Hz) bands [24]. version of the Mini-Mental State Examination (MMSE) [18, 19]. AD patients scored below 26 points. All probands 2.4. Classifiers. The EEG dataset was composed of 1360 did not have a history of diabetes mellitus, kidney disease, epochs (40 epochs of 34 subjects). The analysis was based on thyroid disease, alcoholism, liver disease, lung disease, or the leave-one-subject-out process: 1320 epochs were used for vitamin B12 deficiency to avoid other causes of cognitive training and 40 epochs from one subject for testing. It means impairment. that, each time, the classifier was trained with epochs from all individuals except the one going to be tested. This procedure 2.2. Data Acquisition and Processing. TheEEGswerereco- was performed to test the classifiers discriminative capacity to work with data diverse from that presented in the training rded with 12 bits resolution, band pass of 1–50 Hz, and sampling rate of 200 Hz. A Braintech 3.0 (EMSA “Equipa- period. The leave-one-subject-out process was repeated for mentos M´edicos”) was the recording hardware. Impedance all 34 individuals (34 tests each montage). was maintained below 10 K, and the electrodes were placed according to the International 10–20 System [5, 20]. The 2.4.1. Support Vector Machines (SVMs). SVMs constitute a interconnected ear lobe electrodes reference (without resis- supervised Machine Learning (ML) technique based on the tor) is standard in our laboratory, despite the fact that there Statistical Learning Theory [25]. In this method, a training are controversies regarding which reference is the best [21, dataset (containing known labeled data examples) is used 22]. The EEGs were recorded during 20 minutes. Probands to draw a hyperplane with maximum margin, based on the were awake and relaxed, with closed eyes. Two skilled neu- feature coordinates, which separates the two classes (in our rophysiologists removed EEG artifacts (blinking, drowsiness, case, Controls and AD). Subsequently, the coordinates of muscle movements, or equipment-related artifacts) from the this hyperplane are used to test a dataset and the accuracy recordings. Subsequently, from each EEG, 40 epochs of eight of the model [26]. When classes are not linearly separable, seconds were selected by visual inspection [23]. feature coordinates should be mapped to a higher dimension A 512-point Hamming Fast Fourier Transform (FFT) by a Kernel function. In this new space, the classes become algorithm was used to process the epochs analysis. The linearly separable and the maximum margin hyperplane can windows were 2.5 seconds long with 90% of overlap between then be found [26]. successive windows [23]. EEG signals were filtered using an In this experiment, the Weka tool [27]withdefault infinite impulse response low-pass elliptic filter with a cutoff values was used to the SVM induction. The regularization frequency at 50 Hz and a zero in the frequency of 60 Hz to coefficient of SVM was maintained in C = 1.0, while the eliminate the interference of the power grid (60 Hz). Kernel used was RBF [28]. The cache size was 250007, and the gamma value was 0.01. 2.3. Feature Extraction. Feature extraction is the method used to mining some characteristics of a particular signal 2.4.2. Logistic Regression (LR). Logistic regression is part epoch producing data that can represent events [23]. The of a category of statistical models called generalized linear spectral peak feature (or peak spectrum), chosen in this models. LR is a classification tool frequently used to help work, corresponds to the frequency where the EEG spectrum diagnosis [29]. In this method, the discriminant function amplitude reaches its maximum value. The montages used analyses the sum of the scores of each feature and then were delimitates the boundaries between the two groups [30]. Logistic regression calculates the predicted probability of (i) Biauricular reference (Bar): Fp1-A1, Fp2-A2, F7-A1, different subgroups (in our analysis) falling into a category F8-A2, F3-A1, F4-A2, C3-A1, C-A2, T3-A1, T4-A2, [30]. In LR induction, we also used Weka tool [27]with P3-A1, P4-A2, O1-A1, O2-A2; default values. In this case, the maximum interaction value (ii) Longitudinal Bipolar (Lbp): Fp1-F3, F3-C3, C3-P3, was −1.0, and the ridge value in the log-likelihood was P3-O1, O1-T5, T5-T3, T3-F7, F7-Fp1, Fp2-F4, F4- configured to 1.0. C4, C4-P4, P4-O2, O2-T6, T6-T4, T4-F9, F8-Fp2; (iii) Crossed Bipolar (Bcr): Fp1-Fp2, F7-F3, F3-Fz, Fz-F4, 3. Results and Discussion F4-F8, T3-C3, C3-Cz, Cz-C4, C4-T4, T5-P3, P3-Pz, Table 1 shows the results of both classifiers to each electrode Pz-P4, P4-T6, O1-O2; montage. The columns represent, respectively, from left to (iv) Counterpart bipolar (Bco): F7-F8, F3-F4, T3-T4, C3- right, accuracy, sensitivity (patients correctly diagnosed as C4, P3-P4, T5-T6, O1-O2; AD), and specificity (controls correctly diagnosed as nor- mals). The first line of each montage shows the percentage (v) Cz reference (Czr): Fp1-Cz, Fp2-Cz, F3-Cz, F4-Cz, of epoch classification (mean and standard deviations). F7-Cz, F8-Cz, T3-Cz, T4-Cz, C3-Cz, C4-Cz, T5-Cz, The second line of Table 1 presents the per subject per- T6-Cz, P3-Cz, P4-Cz, O1-Cz, O2-Cz. centage. The leave-one-out analysis of each subject took into Each of these electrode montages (Figure 1)had spectral consideration the ratio between the number of epochs clas- peaks calculated for delta (from 0.1 to 4.0 Hz), theta (from sified correctly and the total number of epochs. When this International Journal of Alzheimer’s Disease 3 Fp1 Fp2 Fp1 Fp2 Fp1 Fp2 F7 F8 F8 F8 F7 F7 Fz Fz Fz F3 F4 F4 F3 F3 F4 A2 A2 A2 A1 A1 A1 T3 T4 T3 T4 T3 T4 Cz Cz Cz C4 C3 C3 C4 C3 C4 P3 Pz P3 Pz P4 P3 P4 P4 T5 T5 Pz T6 T6 T6 T5 O1 O2 O1 O2 O1 O2 Fp1 Fp2 Fp1 Fp2 F7 F8 F8 F7 Fz Fz F3 F4 F3 F4 A2 A2 A1 A1 T3 T4 Cz T3 Cz T4 C4 C4 C3 C3 T5 Pz P4 T6 P3 P3 Pz P4 T6 T5 O1 O1 O2 O2 Figure 1: Spectral peaks montage maps. Lines correspond to subtractions used to calculate spectral peaks. From left to right, top to bottom: Counterpart Bipolar (Bco), Longitudinal Bipolar (Lbp), Crossed Bipolar (Bcr), Biauricular reference (Bar), and Cz reference (Czr). Table 1: Accuracy, sensitivity, and specificity rates for each montage. Best results in bold and worst results in italic. Support Vector Machines Logistic Regression Accuracy (%) Sensitivity (%)Specificity(%)Accuracy(%) Sensitivity (%)Specificity(%) Bipolar Counterpart Bipolar Counterpart Epochs 81,32 ± 28,00 89,43 ± 20,92 66,46 ± 33,84 82,13 ± 20,86 86,93 ± 17,49 73,33±24,32 Patient 85,29 90,91 75,00 91,18 95,45 83,33 Longitudinal Bipolar Longitudinal Bipolar Epochs 72,72 ± 36,80 84,09 ± 27,52 51,88 ± 43,39 66,03 ± 35,76 75,45 ± 31,42 48,75 ± 38,04 Patient 79,41 90,91 58,33 64,71 72,73 50,00 Crossed Bipolar Crossed Bipolar Epochs 69,19 ± 37,60 80,23 ± 32,40 48,96 ± 39,32 65,07 ± 36,09 76,59 ± 33,23 43,96 ± 32,36 Patient 64,71 81,82 41,67 67,65 77,27 50,00 Biauricular Reference Biauricular Reference Epochs 70,07 ± 36,81 85,57 ± 23,12 41,67 ± 41,03 66,32 ± 32,50 76,14 ± 28,05 48,33 ± 33,50 Patient 76,47 95,45 41,67 67,65 81,82 41,67 Cz Reference Cz Reference Epochs 70,22 ± 37,70 81,36 ± 31,15 49,79 ± 41,33 71,62 ± 28,37 80,45 ± 24,64 55,42 ± 28,52 Patient 70,59 81,82 50,00 73,53 86,36 50,00 ratio was over 0.5, the subject classification was considered because they validate this qEEG technique as a diagnostic correct. After 34 tests, the rate of subject correct diagnosis method. Therefore, it can help supporting clinical diagnosis. was calculated. In Table 1, Bco is the montage with highest It is important to note that high standard deviation (SD) number of correct diagnosis and the lowest standard devi- is a methodological consequence of the leave-one-subject- ation to all classifiers. Bco also had high specificity (correct out test. If an individual had bad epochs accuracy, the group diagnosis of AD) and sensibility. These findings are relevant mean was low and the SD high. Bco was the montage with 4 International Journal of Alzheimer’s Disease Table 2: Number of patients with epoch accuracy rates equal to 100%, exceeding or equal to 75%, less than or equal to 50%, and equal to 0% for each test. Best results in bold and worst results in italic. Support Vector Machines Logistic Regression = 100 ≥ 75 ≤ 50 = 0 = 100 ≥ 75 ≤ 50 = 0 Bipolar Counterpart 15 26 5 0 12 25 3 0 Longitudinal Bipolar 16 21 7 2 7 20 12 1 Crossed Bipolar 13 20 12 3 8 18 11 2 Biauricular Reference 14 19 82 8 18 11 0 Cz Reference 15 20 10 34 20 9 0 Table 3: Odds ratio to Bipolar Counterpart LR test. In bold the significant ones (>1). delta theta alpha beta gamma F3-F4 0,371 ± 0,063 2,496 ± 0,575 0,969 ± 0,230 1,188 ± 0,101 0,913 ± 0,052 F7-F8 128,806 ± 50,806 2,580 ± 0,958 0,728 ± 0,221 1,543 ± 0,126 0,659 ± 0,045 C3-C4 0,693 ± 0,176 3,667 ± 1,213 0,734 ± 0,191 2,229 ± 0,275 0,975 ± 0,047 T3-T4 0,753 ± 0,182 0,177 ± 0,068 0,263 ± 0,059 0,836 ± 0,087 1,118 ± 0,049 T5-T6 0,277 ± 0,075 1,011 ± 0,258 0,104 ± 0,029 0,875 ± 0,080 0,649 ± 0,034 P3-P4 0,574 ± 0,107 0,511 ± 0,091 0,019 ± 0,008 0,962 ± 0,118 1,888 ± 0,087 O1-O2 2,231 ± 0.364 0,402 ± 0,126 0,767 ± 0,177 0,761 ± 0,072 0,792 ± 0,036 lower SD, consequently, indicating less variability in number values of ODDR > 1 to beta band, and T3-T4 and P3-P4 of correct diagnosis. presented values of ODDR > 1 to gamma band. Table 2 shows the results of the individual accuracy rate EEGs of mild DA have higher theta activity and low beta variability. The columns show, respectively, from left to right, activity [31, 32], as seen in our cases (F3-F4, F7-F8, C3- epochs accuracy by each subject of 100%, ≥75%, ≤50%, and C4, and T5-T6). Furthermore, these electrodes were directly 0% (all epochs incorrectly classified by one subject). associated with the inter-hemispheric differences found in SVMs tests presented Lbp as the montage with maximum our AD population [33]. Moderate to advanced cases of AD epoch accuracy (16 subjects with 100% accuracy), followed are associated with increasing of delta activity [32, 34–36], by Bcp e Czr (15 cases each). Bco was the montage with and this could explain the values found in F7-F8 and O1- higher number of cases with accuracy greater than or equal O2. Thus, our findings are in accordance to data presented to 75%, less cases with accuracy less than or equal to 50%, by others. and without cases of 0% correct classification. The analysis of the number of electrodes related to The LR tests ratified Bco as having the highest number each montage demonstrates that the montages with higher of 100% accuracy results, the highest number of cases number of signals were Lbp and Czr with 16 signals each, with accuracy greater than or equal to 75%, less cases followed by Bar and Bcr with 14 signals. The montage with with accuracy less than 50%, and no cases of 0% correct lowest number of signals was Bco (7 signals). We can say that classification (in this last case similar to Bar and Czr, both Bco is also the more compact (less electrodes), consequently, with zero cases). less expensive in terms of processing time. This study suggested that Bco was the more trustworthy montage because of his higher rates of 100% epoch accuracy and absence of 0% cases to both classifiers. Consequently, 4. Conclusion other parameters could be tested based on LR. The odds ratio values (ODDR) could be analyzed from the ratio To sum, our results are in accordance with the literature AD/controls (Table 3). It was possible to verify 11 features that suggests that the spectral peak is an efficient tool in AD presenting ODDR > 1. Consequently, there is a possibility diagnosis [24, 37]. Our contribution is to answer the ques- that these features can be associated with AD. tion that gave origin to the paper. Yes, the analysis indicates Among these ODDR features, the electrodes F3-F4, F7- that the bipolar inter-hemispheric montage (Counterpart F8, C3-C4, and T5-T6 presented values of ODDR > 1 to theta bipolar) is the best to evaluate AD patients with the help of band; the electrodes F7-F8 and O1-O2 presented values of automatic classifiers (DA versus N) [38, 39], when using EEG ODDR > 1 to delta band; F3-F4, F7-F8 and C3-C4 presented spectral peaks as features (predictors). International Journal of Alzheimer’s Disease 5 Although more tests are needed to confirm the gener- studies with nicergoline,” Dementia and Geriatric Cognitive Disorders, vol. 8, supplement 1, pp. 12–21, 1997. alization power of our classifiers, we propose that spectral [15] E. Pucci, N. Belardinelli, G. Cacchio, ` M. Signorino, and F. peak calculation using different montages of electrodes have Angeleri, “EEG power spectrum differences in early and late an influence on the classification results (differentiation) of onset forms of Alzheimer’s disease,” Clinical Neurophysiology, normal subjects and patients with AD. Our future goal is vol. 110, no. 4, pp. 621–631, 1999. to generalize the results obtained increasing the number of [16] G. McKhann, D. Drachman, and M. Folstein, “Clinical diag- probands. nosis of Alzheimer’s disease: report of the NINCDS-ADRDA work group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s disease,” References Neurology, vol. 34, no. 7, pp. 939–944, 1984. [1] H. Adeli, S. Ghosh-Dastidar, and N. Dadmehr, “Alzheimer’s [17] American Psychiatric Association, Diagnostic and Statistical Manual (DSM-IV-TR), American Psychiatric Association, 3rd disease: models of computation and analysis of EEGs,” Clinical edition, 2000. EEG and Neuroscience, vol. 36, no. 3, pp. 131–140, 2005. [2] A. Mart´ınez, D. K. Lahiri, E. Giacobini, and N. H. Greig, [18] M. F. Folstein, S. E. Folstein, and P. R. McHugh, “‘Mini mental state’. A practical method for grading the cognitive state of “Advances in alzheimer therapy: understanding pharmacolog- ical approaches to the disease,” Current Alzheimer Research, patients for the clinician,” Journal of Psychiatric Research, vol. 12, no. 3, pp. 189–198, 1975. vol. 6, no. 2, pp. 83–85, 2009. [3] R. Sneddon, W. R. Shankle, J. Hara, A. Rodriquez, D. [19] S. M. D. Brucki, R. Nitrin, P. Caramelli, P. H. F. Bertolucci, and I. H. Okamoto, “Suggestions for utilization of the mini-mental Hoffman, and U. Saha, “qEEG monitoring of Alzheimer’s disease treatment: a preliminary report of three case studies,” state examination in Brazil,” Arquivos de Neuro-Psiquiatria, vol. 61, no. 3B, pp. 777–781, 2003. Clinical EEG and Neuroscience, vol. 37, no. 1, pp. 54–59, 2006. [4] M. R. Nuwer, G. Comi, R. Emerson et al., “IFCN standards for [20] G. H. Klem,H.O.Luders, ¨ H. H. Jasper, and C. Elger, “The ten-twenty electrode system of the International Federation. digital recording of clinical EEG,” Electroencephalography and Clinical Neurophysiology, vol. 106, no. 3, pp. 259–261, 1998. The International Federation of Clinical Neurophysiology,” Electroencephalography and Clinical Neurophysiology, vol. 52, [5] F. J. C. Luccas, R. Anghinah, N. I. O. Braga et al., “Guidelines pp. 3–6, 1999. for recording/analyzing quantitative EEG and evoked poten- tials: part II. Clinical aspects,” Arquivos de Neuro-Psiquiatria, [21] P. Rappelsberger, “The reference problem and mapping of coherence: a simulation study,” Brain Topography, vol. 2, no. vol. 57, no. 1, pp. 132–146, 1999. [6] D.W.Klass andR.P.Brenner, “Electroencephalographyofthe 1-2, pp. 63–72, 1989. [22] D. Yao, L. Wang, L. Arendt-Nielsen, and A. C. Chen, “The elderly,” Journal of Clinical Neurophysiology,vol. 12, no.2,pp. 116–131, 1995. effect of reference choices on the spatio-temporal analysis of brain evoked potentials: the use of infinite reference,” [7] M. C. Sandmann, E. R. Piana, D. S. Sousa, and P. R. M. De Bittencourt, “Digital EEG with brain mapping in Alzheimer’s Computers in Biology and Medicine, vol. 37, no. 11, pp. 1529– 1538, 2007. dementia and Parkinson’s disease: a prospective controlled study,” Arquivos de Neuro-Psiquiatria, vol. 54, no. 1, pp. 50–56, [23] S. Sanei and J. Chambers, EEG Signal Processing, Wiley- Interscience, New York, NY, USA, 2007. [8] D. Lehmann, “Multichannel topography of human alpha EEG [24] R. Anghinah, Estudo da densidade espectral e da coerˆencia do fields,” Electroencephalography and Clinical Neurophysiology, eletroencefalograma em indiv´ıduos normais e com doen¸a c de Alzheimer provavel, Ph.D. thesis, Faculdade de Medicina da vol. 31, no. 5, pp. 439–449, 1971. ´ [9] F.H.Duffy, J. L. Burchfiel, and C. T. Lombroso, “Brain Universidade de Sao ˜ Paulo, Sao ˜ Paulo, Brazil, 2003. [25] V. N. Vapnik, The Nature of Statistical Learning Theory, electrical activity mapping (BEAM): a method for extending the clinical utility of EEG and evoked potential data,” Annals Springer, New York, NY, USA, 1995. [26] N. Cristianini and J. Shawe-Taylor, An Introduction to Support of Neurology, vol. 5, no. 4, pp. 309–321, 1979. [10] M. Martin-Loeches, P. Gil, F. Jimenez et al., “Topographic Vector Machines: and Other Kernel-Based Learning Methods, Cambridge University Press, Cambridge, UK, 2000. maps of brain electrical activity in primary degenerative dementia of the Alzheimer type and multiinfarct dementia,” [27] I. A. Witten and E. Frank, Data Mining: Practical Machine Learning Tools and Techniques, Elsevier, 2005. Biological Psychiatry, vol. 29, no. 3, pp. 211–223, 1991. [11] T. Dierks, I. Perisic, L. Frolich, R. Ihl, and K. Maurer, “Topog- [28] L. R. Trambaiolli, A. C. Lorena, F. J. Fraga, and R. Anghinah, “Uso de aprendizado de maquina ´ no aux´ılio ao diagnostic ´ o de raphy of the quantitative electroencephalogram in dementia of the Alzheimer type: relation to severity of dementia,” Alzheimer,” REIC—Revista Eletronica ˆ de Iniciac¸ao ˜ Cient´ıfica, Psychiatry Research, vol. 40, no. 3, pp. 181–194, 1991. vol. 9, no. 3, pp. 1–14, 2009 (Portuguese). [29] D. Pregibon, “Logistic regression diagnostics,” The Annals of [12] A. F. Leuchter, I. A. Cook, T. F. Newton et al., “Regional differ- ences in brain electrical activity in dementia: use of spectral Statistics, vol. 9, no. 4, pp. 705–724, 1981. [30] D. Hosmer and S. Lemeshow, Applied Logistic Regression, power and spectral ratio measures,” Electroencephalography and Clinical Neurophysiology, vol. 87, no. 6, pp. 385–393, 1993. Wiley-Interscience, New York, NY, USA, 2000. [31] L. A. Coben, W. L. Danziger, and L. Berg, “Frequency analysis [13] A. F. Leuchter, J. E. Spar, D. O. Walter, and H. Weiner, “Elec- troencephalographic spectra and coherence in the diagnosis of of the resting awake EEG in mild senile dementia of Alzheimer type,” Electroencephalography and Clinical Neurophysiology, Alzheimer’s-type and multi-infarct dementia. A pilot study,” Archives of General Psychiatry, vol. 44, no. 11, pp. 993–998, vol. 55, no. 4, pp. 372–380, 1983. [32] L. A. Coben, W. Danziger, and M. Storandt, “A longitudinal [14] B. Saletu, P. Anderer, and H. V. Semlitsch, “Relations between EEG study of mild senile dementia of Alzheimer type: changes at 1 year and at 2.5 years,” Electroencephalography and Clinical symptomatology and brain function in dementias: double- blind, placebo-controlled, clinical and EEG/ERP mapping Neurophysiology, vol. 61, no. 2, pp. 101–112, 1985. 6 International Journal of Alzheimer’s Disease [33] F. H. Duffy, M. S. Albert, and G. McAnulty, “Brain electrical activity in patients with presenile and senile dementia of the Alzheimer type,” Annals of Neurology, vol. 16, no. 4, pp. 439– 448, 1984. [34] D. B. Hier, C. A. Mangone, R. Ganellen et al., “Quantitative measurement of delta activity in Alzheimer’s disease,” Clinical EEG Electroencephalography, vol. 22, no. 3, pp. 178–182, 1991. [35] M. Penttila, ¨ J. V. Partanen, H. Soininen, and P. J. Riekkinen, “Quantitative analysis of occipital EEG in different stages of Alzheimer’s disease,” Electroencephalography and Clinical Neurophysiology, vol. 60, no. 1, pp. 1–6, 1985. [36] B. Stigsby, G. Johannesson, and D. H. Ingvar, “Regional EEG analysis and regional cerebral blood flow in Alzheimer’s and Pick’s diseases,” Electroencephalography and Clinical Neuro- physiology, vol. 51, no. 5, pp. 537–547, 1981. [37] I. Raicher, D. Takahashi, P. Kanda, R. Nitrini, and R. Anghinah, “qEEG spectral peak in Alzheimer’s disease,” Dementia and Neuropsychologia, vol. 2, no. 1, pp. 9–12, 2008. [38] L. R. Trambaiolli, A. C. Lorena, F. J. Fraga, and R. Anghinah, “Support vector machines in the diagnosis of Alzheimer’s disease,” in Proceedings of the ISSNIP Biosignals and Biorobotics Conference, vol. 1, pp. 1–6, 2010. [39] Y. T. Kwak, “Quantitative EEG findings in different stages of Alzheimer’s disease,” Journal of Clinical Neurophysiologyy, vol. 23, no. 5, pp. 456–461, 2006. MEDIATORS of INFLAMMATION The Scientific Gastroenterology Journal of World Journal Research and Practice Diabetes Research Disease Markers Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 International Journal of Journal of Immunology Research Endocrinology Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 Submit your manuscripts at http://www.hindawi.com BioMed PPAR Research Research International Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 Journal of Obesity Evidence-Based Journal of Journal of Stem Cells Complementary and Ophthalmology International Alternative Medicine Oncology Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 Parkinson’s Disease Computational and Behavioural Mathematical Methods AIDS Oxidative Medicine and in Medicine Research and Treatment Cellular Longevity Neurology Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Alzheimer's Disease Hindawi Publishing Corporation

Does EEG Montage Influence Alzheimer's Disease Electroclinic Diagnosis?

Loading next page...
 
/lp/hindawi-publishing-corporation/does-eeg-montage-influence-alzheimer-s-disease-electroclinic-diagnosis-0gvKNYLC2P

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Hindawi Publishing Corporation
Copyright
Copyright © 2011 L. R. Trambaiolli et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ISSN
2090-8024
DOI
10.4061/2011/761891
Publisher site
See Article on Publisher Site

Abstract

SAGE-Hindawi Access to Research International Journal of Alzheimer’s Disease Volume 2011, Article ID 761891, 6 pages doi:10.4061/2011/761891 Research Article Does EEG Montage Influence Alzheimer’s Disease Electroclinic Diagnosis? 1 1 2 3 3 3 L. R. Trambaiolli, A. C. Lorena, F. J. Fraga, P. A. M. K. Kanda, R. Nitrini, and R. Anghinah Mathematics, Computing and Cognition Center (CMCC), Universidade Federal do ABC (UFABC), Rua Santa Adelia, 166, 09210-170 Santo Andre, SP, Brazil Engineering, Modeling and Applied Social Sciences Center (CECS), Universidade Federal do ABC (UFABC), Rua Santa Adelia, 166, 09210-170 Santo Andre, SP, Brazil Reference Center of Behavioral Disturbances and Dementia (CEREDIC) and Neurology, Department of Medicine School of University of Sao ˜ Paulo (USP), Rua Arruda Alvim, 206, 05.410-020 Sao ˜ Paulo, SP, Brazil Correspondence should be addressed to F. J. Fraga, franciscojfraga@gmail.com Received 23 December 2010; Revised 23 February 2011; Accepted 7 March 2011 Academic Editor: Fabrizio Vecchio Copyright © 2011 L. R. Trambaiolli et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. There is not a specific Alzheimer’s disease (AD) diagnostic test. AD diagnosis relies on clinical history, neuropsychological, and laboratory tests, neuroimaging and electroencephalography. Therefore, new approaches are necessary to enable earlier and more accurate diagnosis and to measure treatment results. Quantitative EEG (qEEG) can be used as a diagnostic tool in selected cases. The aim of this study was to answer if distinct electrode montages have different sensitivity when differentiating controls from AD patients. We analyzed EEG spectral peaks (delta, theta, alpha, beta, and gamma bands), and we compared references (Biauricular, Longitudinal bipolar, Crossed bipolar, Counterpart bipolar, and Cz reference). Support Vector Machines and Logistic Regression classifiers showed Counterpart bipolar montage as the most sensitive electrode combination. Our results suggest that Counterpart bipolar montage is the best choice to study EEG spectral peaks of controls versus AD. 1. Introduction [10–13] in AD EEG background. Saletu et al. [14]found a localized temporal decrease of alpha and beta activities in Alzheimer’s disease (AD) diagnosis is based upon clinical AD and slow cerebral rhythms widespread distribution in history, neuropsychological and laboratory tests, neuroimag- vascular dementia (VaD) [10–13]. Pucci et al. [15] proposed that a decrease in alpha frequency to 6.0–8.0 Hz could be an ing, and electroencephalography (EEG). New approaches are necessary to earlier and more accurate diagnosis [1, 2]and to AD marker. measure treatment results [3]. Despite the knowledge grounded in this field during the last decades, there are lots of unanswered questions that hin- EEG visual analysis can be a helpful diagnostic test in AD [4, 5]. Background frequency displacement to delta and theta der qEEG consolidation as an AD diagnostic tool. Our objec- frequencies and the dropout of central alpha rhythm are tive was to study if distinct electrode montages have different sensitivity when differentiating controls from AD patients. common EEG findings in AD [6]. Accordingly, Sandmann et al. [7] observed a direct correlation between the degree of cognitive impairment and the power of low-frequency 2. Materials and Methods electrical activity in the EEG. Since the first quantitative EEG (qEEG) studies by 2.1. Subjects. The dataset was composed of electroencep- Lehmann [8]and Duffy et al. [9], spectral analysis (specA) halograms (EEGs) recorded from two groups aged from 60 and statistics have been applied to EEG. Moreover, specA to 80 years: (S1) 12 normal subjects and (S2) 22 probable has been considered from 71% to 81% sensitive to changes AD patients (NINCDS-ADRDA criteria) [16]. AD group 2 International Journal of Alzheimer’s Disease was classified as having mild to moderate symptoms (DSM- 4.0 to 8.0 Hz), alpha (from 8.0 to 12.0 Hz), beta (from 12.0 IV-TR) [17]. Both groups were submitted to the Brazilian to 30.0 Hz), and gamma (from 30.0 to 50.0 Hz) bands [24]. version of the Mini-Mental State Examination (MMSE) [18, 19]. AD patients scored below 26 points. All probands 2.4. Classifiers. The EEG dataset was composed of 1360 did not have a history of diabetes mellitus, kidney disease, epochs (40 epochs of 34 subjects). The analysis was based on thyroid disease, alcoholism, liver disease, lung disease, or the leave-one-subject-out process: 1320 epochs were used for vitamin B12 deficiency to avoid other causes of cognitive training and 40 epochs from one subject for testing. It means impairment. that, each time, the classifier was trained with epochs from all individuals except the one going to be tested. This procedure 2.2. Data Acquisition and Processing. TheEEGswerereco- was performed to test the classifiers discriminative capacity to work with data diverse from that presented in the training rded with 12 bits resolution, band pass of 1–50 Hz, and sampling rate of 200 Hz. A Braintech 3.0 (EMSA “Equipa- period. The leave-one-subject-out process was repeated for mentos M´edicos”) was the recording hardware. Impedance all 34 individuals (34 tests each montage). was maintained below 10 K, and the electrodes were placed according to the International 10–20 System [5, 20]. The 2.4.1. Support Vector Machines (SVMs). SVMs constitute a interconnected ear lobe electrodes reference (without resis- supervised Machine Learning (ML) technique based on the tor) is standard in our laboratory, despite the fact that there Statistical Learning Theory [25]. In this method, a training are controversies regarding which reference is the best [21, dataset (containing known labeled data examples) is used 22]. The EEGs were recorded during 20 minutes. Probands to draw a hyperplane with maximum margin, based on the were awake and relaxed, with closed eyes. Two skilled neu- feature coordinates, which separates the two classes (in our rophysiologists removed EEG artifacts (blinking, drowsiness, case, Controls and AD). Subsequently, the coordinates of muscle movements, or equipment-related artifacts) from the this hyperplane are used to test a dataset and the accuracy recordings. Subsequently, from each EEG, 40 epochs of eight of the model [26]. When classes are not linearly separable, seconds were selected by visual inspection [23]. feature coordinates should be mapped to a higher dimension A 512-point Hamming Fast Fourier Transform (FFT) by a Kernel function. In this new space, the classes become algorithm was used to process the epochs analysis. The linearly separable and the maximum margin hyperplane can windows were 2.5 seconds long with 90% of overlap between then be found [26]. successive windows [23]. EEG signals were filtered using an In this experiment, the Weka tool [27]withdefault infinite impulse response low-pass elliptic filter with a cutoff values was used to the SVM induction. The regularization frequency at 50 Hz and a zero in the frequency of 60 Hz to coefficient of SVM was maintained in C = 1.0, while the eliminate the interference of the power grid (60 Hz). Kernel used was RBF [28]. The cache size was 250007, and the gamma value was 0.01. 2.3. Feature Extraction. Feature extraction is the method used to mining some characteristics of a particular signal 2.4.2. Logistic Regression (LR). Logistic regression is part epoch producing data that can represent events [23]. The of a category of statistical models called generalized linear spectral peak feature (or peak spectrum), chosen in this models. LR is a classification tool frequently used to help work, corresponds to the frequency where the EEG spectrum diagnosis [29]. In this method, the discriminant function amplitude reaches its maximum value. The montages used analyses the sum of the scores of each feature and then were delimitates the boundaries between the two groups [30]. Logistic regression calculates the predicted probability of (i) Biauricular reference (Bar): Fp1-A1, Fp2-A2, F7-A1, different subgroups (in our analysis) falling into a category F8-A2, F3-A1, F4-A2, C3-A1, C-A2, T3-A1, T4-A2, [30]. In LR induction, we also used Weka tool [27]with P3-A1, P4-A2, O1-A1, O2-A2; default values. In this case, the maximum interaction value (ii) Longitudinal Bipolar (Lbp): Fp1-F3, F3-C3, C3-P3, was −1.0, and the ridge value in the log-likelihood was P3-O1, O1-T5, T5-T3, T3-F7, F7-Fp1, Fp2-F4, F4- configured to 1.0. C4, C4-P4, P4-O2, O2-T6, T6-T4, T4-F9, F8-Fp2; (iii) Crossed Bipolar (Bcr): Fp1-Fp2, F7-F3, F3-Fz, Fz-F4, 3. Results and Discussion F4-F8, T3-C3, C3-Cz, Cz-C4, C4-T4, T5-P3, P3-Pz, Table 1 shows the results of both classifiers to each electrode Pz-P4, P4-T6, O1-O2; montage. The columns represent, respectively, from left to (iv) Counterpart bipolar (Bco): F7-F8, F3-F4, T3-T4, C3- right, accuracy, sensitivity (patients correctly diagnosed as C4, P3-P4, T5-T6, O1-O2; AD), and specificity (controls correctly diagnosed as nor- mals). The first line of each montage shows the percentage (v) Cz reference (Czr): Fp1-Cz, Fp2-Cz, F3-Cz, F4-Cz, of epoch classification (mean and standard deviations). F7-Cz, F8-Cz, T3-Cz, T4-Cz, C3-Cz, C4-Cz, T5-Cz, The second line of Table 1 presents the per subject per- T6-Cz, P3-Cz, P4-Cz, O1-Cz, O2-Cz. centage. The leave-one-out analysis of each subject took into Each of these electrode montages (Figure 1)had spectral consideration the ratio between the number of epochs clas- peaks calculated for delta (from 0.1 to 4.0 Hz), theta (from sified correctly and the total number of epochs. When this International Journal of Alzheimer’s Disease 3 Fp1 Fp2 Fp1 Fp2 Fp1 Fp2 F7 F8 F8 F8 F7 F7 Fz Fz Fz F3 F4 F4 F3 F3 F4 A2 A2 A2 A1 A1 A1 T3 T4 T3 T4 T3 T4 Cz Cz Cz C4 C3 C3 C4 C3 C4 P3 Pz P3 Pz P4 P3 P4 P4 T5 T5 Pz T6 T6 T6 T5 O1 O2 O1 O2 O1 O2 Fp1 Fp2 Fp1 Fp2 F7 F8 F8 F7 Fz Fz F3 F4 F3 F4 A2 A2 A1 A1 T3 T4 Cz T3 Cz T4 C4 C4 C3 C3 T5 Pz P4 T6 P3 P3 Pz P4 T6 T5 O1 O1 O2 O2 Figure 1: Spectral peaks montage maps. Lines correspond to subtractions used to calculate spectral peaks. From left to right, top to bottom: Counterpart Bipolar (Bco), Longitudinal Bipolar (Lbp), Crossed Bipolar (Bcr), Biauricular reference (Bar), and Cz reference (Czr). Table 1: Accuracy, sensitivity, and specificity rates for each montage. Best results in bold and worst results in italic. Support Vector Machines Logistic Regression Accuracy (%) Sensitivity (%)Specificity(%)Accuracy(%) Sensitivity (%)Specificity(%) Bipolar Counterpart Bipolar Counterpart Epochs 81,32 ± 28,00 89,43 ± 20,92 66,46 ± 33,84 82,13 ± 20,86 86,93 ± 17,49 73,33±24,32 Patient 85,29 90,91 75,00 91,18 95,45 83,33 Longitudinal Bipolar Longitudinal Bipolar Epochs 72,72 ± 36,80 84,09 ± 27,52 51,88 ± 43,39 66,03 ± 35,76 75,45 ± 31,42 48,75 ± 38,04 Patient 79,41 90,91 58,33 64,71 72,73 50,00 Crossed Bipolar Crossed Bipolar Epochs 69,19 ± 37,60 80,23 ± 32,40 48,96 ± 39,32 65,07 ± 36,09 76,59 ± 33,23 43,96 ± 32,36 Patient 64,71 81,82 41,67 67,65 77,27 50,00 Biauricular Reference Biauricular Reference Epochs 70,07 ± 36,81 85,57 ± 23,12 41,67 ± 41,03 66,32 ± 32,50 76,14 ± 28,05 48,33 ± 33,50 Patient 76,47 95,45 41,67 67,65 81,82 41,67 Cz Reference Cz Reference Epochs 70,22 ± 37,70 81,36 ± 31,15 49,79 ± 41,33 71,62 ± 28,37 80,45 ± 24,64 55,42 ± 28,52 Patient 70,59 81,82 50,00 73,53 86,36 50,00 ratio was over 0.5, the subject classification was considered because they validate this qEEG technique as a diagnostic correct. After 34 tests, the rate of subject correct diagnosis method. Therefore, it can help supporting clinical diagnosis. was calculated. In Table 1, Bco is the montage with highest It is important to note that high standard deviation (SD) number of correct diagnosis and the lowest standard devi- is a methodological consequence of the leave-one-subject- ation to all classifiers. Bco also had high specificity (correct out test. If an individual had bad epochs accuracy, the group diagnosis of AD) and sensibility. These findings are relevant mean was low and the SD high. Bco was the montage with 4 International Journal of Alzheimer’s Disease Table 2: Number of patients with epoch accuracy rates equal to 100%, exceeding or equal to 75%, less than or equal to 50%, and equal to 0% for each test. Best results in bold and worst results in italic. Support Vector Machines Logistic Regression = 100 ≥ 75 ≤ 50 = 0 = 100 ≥ 75 ≤ 50 = 0 Bipolar Counterpart 15 26 5 0 12 25 3 0 Longitudinal Bipolar 16 21 7 2 7 20 12 1 Crossed Bipolar 13 20 12 3 8 18 11 2 Biauricular Reference 14 19 82 8 18 11 0 Cz Reference 15 20 10 34 20 9 0 Table 3: Odds ratio to Bipolar Counterpart LR test. In bold the significant ones (>1). delta theta alpha beta gamma F3-F4 0,371 ± 0,063 2,496 ± 0,575 0,969 ± 0,230 1,188 ± 0,101 0,913 ± 0,052 F7-F8 128,806 ± 50,806 2,580 ± 0,958 0,728 ± 0,221 1,543 ± 0,126 0,659 ± 0,045 C3-C4 0,693 ± 0,176 3,667 ± 1,213 0,734 ± 0,191 2,229 ± 0,275 0,975 ± 0,047 T3-T4 0,753 ± 0,182 0,177 ± 0,068 0,263 ± 0,059 0,836 ± 0,087 1,118 ± 0,049 T5-T6 0,277 ± 0,075 1,011 ± 0,258 0,104 ± 0,029 0,875 ± 0,080 0,649 ± 0,034 P3-P4 0,574 ± 0,107 0,511 ± 0,091 0,019 ± 0,008 0,962 ± 0,118 1,888 ± 0,087 O1-O2 2,231 ± 0.364 0,402 ± 0,126 0,767 ± 0,177 0,761 ± 0,072 0,792 ± 0,036 lower SD, consequently, indicating less variability in number values of ODDR > 1 to beta band, and T3-T4 and P3-P4 of correct diagnosis. presented values of ODDR > 1 to gamma band. Table 2 shows the results of the individual accuracy rate EEGs of mild DA have higher theta activity and low beta variability. The columns show, respectively, from left to right, activity [31, 32], as seen in our cases (F3-F4, F7-F8, C3- epochs accuracy by each subject of 100%, ≥75%, ≤50%, and C4, and T5-T6). Furthermore, these electrodes were directly 0% (all epochs incorrectly classified by one subject). associated with the inter-hemispheric differences found in SVMs tests presented Lbp as the montage with maximum our AD population [33]. Moderate to advanced cases of AD epoch accuracy (16 subjects with 100% accuracy), followed are associated with increasing of delta activity [32, 34–36], by Bcp e Czr (15 cases each). Bco was the montage with and this could explain the values found in F7-F8 and O1- higher number of cases with accuracy greater than or equal O2. Thus, our findings are in accordance to data presented to 75%, less cases with accuracy less than or equal to 50%, by others. and without cases of 0% correct classification. The analysis of the number of electrodes related to The LR tests ratified Bco as having the highest number each montage demonstrates that the montages with higher of 100% accuracy results, the highest number of cases number of signals were Lbp and Czr with 16 signals each, with accuracy greater than or equal to 75%, less cases followed by Bar and Bcr with 14 signals. The montage with with accuracy less than 50%, and no cases of 0% correct lowest number of signals was Bco (7 signals). We can say that classification (in this last case similar to Bar and Czr, both Bco is also the more compact (less electrodes), consequently, with zero cases). less expensive in terms of processing time. This study suggested that Bco was the more trustworthy montage because of his higher rates of 100% epoch accuracy and absence of 0% cases to both classifiers. Consequently, 4. Conclusion other parameters could be tested based on LR. The odds ratio values (ODDR) could be analyzed from the ratio To sum, our results are in accordance with the literature AD/controls (Table 3). It was possible to verify 11 features that suggests that the spectral peak is an efficient tool in AD presenting ODDR > 1. Consequently, there is a possibility diagnosis [24, 37]. Our contribution is to answer the ques- that these features can be associated with AD. tion that gave origin to the paper. Yes, the analysis indicates Among these ODDR features, the electrodes F3-F4, F7- that the bipolar inter-hemispheric montage (Counterpart F8, C3-C4, and T5-T6 presented values of ODDR > 1 to theta bipolar) is the best to evaluate AD patients with the help of band; the electrodes F7-F8 and O1-O2 presented values of automatic classifiers (DA versus N) [38, 39], when using EEG ODDR > 1 to delta band; F3-F4, F7-F8 and C3-C4 presented spectral peaks as features (predictors). International Journal of Alzheimer’s Disease 5 Although more tests are needed to confirm the gener- studies with nicergoline,” Dementia and Geriatric Cognitive Disorders, vol. 8, supplement 1, pp. 12–21, 1997. alization power of our classifiers, we propose that spectral [15] E. Pucci, N. Belardinelli, G. Cacchio, ` M. Signorino, and F. peak calculation using different montages of electrodes have Angeleri, “EEG power spectrum differences in early and late an influence on the classification results (differentiation) of onset forms of Alzheimer’s disease,” Clinical Neurophysiology, normal subjects and patients with AD. Our future goal is vol. 110, no. 4, pp. 621–631, 1999. to generalize the results obtained increasing the number of [16] G. McKhann, D. Drachman, and M. Folstein, “Clinical diag- probands. nosis of Alzheimer’s disease: report of the NINCDS-ADRDA work group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s disease,” References Neurology, vol. 34, no. 7, pp. 939–944, 1984. [1] H. Adeli, S. Ghosh-Dastidar, and N. Dadmehr, “Alzheimer’s [17] American Psychiatric Association, Diagnostic and Statistical Manual (DSM-IV-TR), American Psychiatric Association, 3rd disease: models of computation and analysis of EEGs,” Clinical edition, 2000. EEG and Neuroscience, vol. 36, no. 3, pp. 131–140, 2005. [2] A. Mart´ınez, D. K. Lahiri, E. Giacobini, and N. H. Greig, [18] M. F. Folstein, S. E. Folstein, and P. R. McHugh, “‘Mini mental state’. A practical method for grading the cognitive state of “Advances in alzheimer therapy: understanding pharmacolog- ical approaches to the disease,” Current Alzheimer Research, patients for the clinician,” Journal of Psychiatric Research, vol. 12, no. 3, pp. 189–198, 1975. vol. 6, no. 2, pp. 83–85, 2009. [3] R. Sneddon, W. R. Shankle, J. Hara, A. Rodriquez, D. [19] S. M. D. Brucki, R. Nitrin, P. Caramelli, P. H. F. Bertolucci, and I. H. Okamoto, “Suggestions for utilization of the mini-mental Hoffman, and U. Saha, “qEEG monitoring of Alzheimer’s disease treatment: a preliminary report of three case studies,” state examination in Brazil,” Arquivos de Neuro-Psiquiatria, vol. 61, no. 3B, pp. 777–781, 2003. Clinical EEG and Neuroscience, vol. 37, no. 1, pp. 54–59, 2006. [4] M. R. Nuwer, G. Comi, R. Emerson et al., “IFCN standards for [20] G. H. Klem,H.O.Luders, ¨ H. H. Jasper, and C. Elger, “The ten-twenty electrode system of the International Federation. digital recording of clinical EEG,” Electroencephalography and Clinical Neurophysiology, vol. 106, no. 3, pp. 259–261, 1998. The International Federation of Clinical Neurophysiology,” Electroencephalography and Clinical Neurophysiology, vol. 52, [5] F. J. C. Luccas, R. Anghinah, N. I. O. Braga et al., “Guidelines pp. 3–6, 1999. for recording/analyzing quantitative EEG and evoked poten- tials: part II. Clinical aspects,” Arquivos de Neuro-Psiquiatria, [21] P. Rappelsberger, “The reference problem and mapping of coherence: a simulation study,” Brain Topography, vol. 2, no. vol. 57, no. 1, pp. 132–146, 1999. [6] D.W.Klass andR.P.Brenner, “Electroencephalographyofthe 1-2, pp. 63–72, 1989. [22] D. Yao, L. Wang, L. Arendt-Nielsen, and A. C. Chen, “The elderly,” Journal of Clinical Neurophysiology,vol. 12, no.2,pp. 116–131, 1995. effect of reference choices on the spatio-temporal analysis of brain evoked potentials: the use of infinite reference,” [7] M. C. Sandmann, E. R. Piana, D. S. Sousa, and P. R. M. De Bittencourt, “Digital EEG with brain mapping in Alzheimer’s Computers in Biology and Medicine, vol. 37, no. 11, pp. 1529– 1538, 2007. dementia and Parkinson’s disease: a prospective controlled study,” Arquivos de Neuro-Psiquiatria, vol. 54, no. 1, pp. 50–56, [23] S. Sanei and J. Chambers, EEG Signal Processing, Wiley- Interscience, New York, NY, USA, 2007. [8] D. Lehmann, “Multichannel topography of human alpha EEG [24] R. Anghinah, Estudo da densidade espectral e da coerˆencia do fields,” Electroencephalography and Clinical Neurophysiology, eletroencefalograma em indiv´ıduos normais e com doen¸a c de Alzheimer provavel, Ph.D. thesis, Faculdade de Medicina da vol. 31, no. 5, pp. 439–449, 1971. ´ [9] F.H.Duffy, J. L. Burchfiel, and C. T. Lombroso, “Brain Universidade de Sao ˜ Paulo, Sao ˜ Paulo, Brazil, 2003. [25] V. N. Vapnik, The Nature of Statistical Learning Theory, electrical activity mapping (BEAM): a method for extending the clinical utility of EEG and evoked potential data,” Annals Springer, New York, NY, USA, 1995. [26] N. Cristianini and J. Shawe-Taylor, An Introduction to Support of Neurology, vol. 5, no. 4, pp. 309–321, 1979. [10] M. Martin-Loeches, P. Gil, F. Jimenez et al., “Topographic Vector Machines: and Other Kernel-Based Learning Methods, Cambridge University Press, Cambridge, UK, 2000. maps of brain electrical activity in primary degenerative dementia of the Alzheimer type and multiinfarct dementia,” [27] I. A. Witten and E. Frank, Data Mining: Practical Machine Learning Tools and Techniques, Elsevier, 2005. Biological Psychiatry, vol. 29, no. 3, pp. 211–223, 1991. [11] T. Dierks, I. Perisic, L. Frolich, R. Ihl, and K. Maurer, “Topog- [28] L. R. Trambaiolli, A. C. Lorena, F. J. Fraga, and R. Anghinah, “Uso de aprendizado de maquina ´ no aux´ılio ao diagnostic ´ o de raphy of the quantitative electroencephalogram in dementia of the Alzheimer type: relation to severity of dementia,” Alzheimer,” REIC—Revista Eletronica ˆ de Iniciac¸ao ˜ Cient´ıfica, Psychiatry Research, vol. 40, no. 3, pp. 181–194, 1991. vol. 9, no. 3, pp. 1–14, 2009 (Portuguese). [29] D. Pregibon, “Logistic regression diagnostics,” The Annals of [12] A. F. Leuchter, I. A. Cook, T. F. Newton et al., “Regional differ- ences in brain electrical activity in dementia: use of spectral Statistics, vol. 9, no. 4, pp. 705–724, 1981. [30] D. Hosmer and S. Lemeshow, Applied Logistic Regression, power and spectral ratio measures,” Electroencephalography and Clinical Neurophysiology, vol. 87, no. 6, pp. 385–393, 1993. Wiley-Interscience, New York, NY, USA, 2000. [31] L. A. Coben, W. L. Danziger, and L. Berg, “Frequency analysis [13] A. F. Leuchter, J. E. Spar, D. O. Walter, and H. Weiner, “Elec- troencephalographic spectra and coherence in the diagnosis of of the resting awake EEG in mild senile dementia of Alzheimer type,” Electroencephalography and Clinical Neurophysiology, Alzheimer’s-type and multi-infarct dementia. A pilot study,” Archives of General Psychiatry, vol. 44, no. 11, pp. 993–998, vol. 55, no. 4, pp. 372–380, 1983. [32] L. A. Coben, W. Danziger, and M. Storandt, “A longitudinal [14] B. Saletu, P. Anderer, and H. V. Semlitsch, “Relations between EEG study of mild senile dementia of Alzheimer type: changes at 1 year and at 2.5 years,” Electroencephalography and Clinical symptomatology and brain function in dementias: double- blind, placebo-controlled, clinical and EEG/ERP mapping Neurophysiology, vol. 61, no. 2, pp. 101–112, 1985. 6 International Journal of Alzheimer’s Disease [33] F. H. Duffy, M. S. Albert, and G. McAnulty, “Brain electrical activity in patients with presenile and senile dementia of the Alzheimer type,” Annals of Neurology, vol. 16, no. 4, pp. 439– 448, 1984. [34] D. B. Hier, C. A. Mangone, R. Ganellen et al., “Quantitative measurement of delta activity in Alzheimer’s disease,” Clinical EEG Electroencephalography, vol. 22, no. 3, pp. 178–182, 1991. [35] M. Penttila, ¨ J. V. Partanen, H. Soininen, and P. J. Riekkinen, “Quantitative analysis of occipital EEG in different stages of Alzheimer’s disease,” Electroencephalography and Clinical Neurophysiology, vol. 60, no. 1, pp. 1–6, 1985. [36] B. Stigsby, G. Johannesson, and D. H. Ingvar, “Regional EEG analysis and regional cerebral blood flow in Alzheimer’s and Pick’s diseases,” Electroencephalography and Clinical Neuro- physiology, vol. 51, no. 5, pp. 537–547, 1981. [37] I. Raicher, D. Takahashi, P. Kanda, R. Nitrini, and R. Anghinah, “qEEG spectral peak in Alzheimer’s disease,” Dementia and Neuropsychologia, vol. 2, no. 1, pp. 9–12, 2008. [38] L. R. Trambaiolli, A. C. Lorena, F. J. Fraga, and R. Anghinah, “Support vector machines in the diagnosis of Alzheimer’s disease,” in Proceedings of the ISSNIP Biosignals and Biorobotics Conference, vol. 1, pp. 1–6, 2010. [39] Y. T. Kwak, “Quantitative EEG findings in different stages of Alzheimer’s disease,” Journal of Clinical Neurophysiologyy, vol. 23, no. 5, pp. 456–461, 2006. MEDIATORS of INFLAMMATION The Scientific Gastroenterology Journal of World Journal Research and Practice Diabetes Research Disease Markers Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 International Journal of Journal of Immunology Research Endocrinology Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 Submit your manuscripts at http://www.hindawi.com BioMed PPAR Research Research International Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 Journal of Obesity Evidence-Based Journal of Journal of Stem Cells Complementary and Ophthalmology International Alternative Medicine Oncology Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 Parkinson’s Disease Computational and Behavioural Mathematical Methods AIDS Oxidative Medicine and in Medicine Research and Treatment Cellular Longevity Neurology Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal

International Journal of Alzheimer's DiseaseHindawi Publishing Corporation

Published: May 2, 2011

References