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

Learn More →

Cognitive robots in the development and rehabilitation of children with developmental disorders

Cognitive robots in the development and rehabilitation of children with developmental disorders Cognitive robots constitute a highly interdisciplinary approach to the issue of therapy of children with developmental disorders. Cognitive robots become more popular, especially in action and language integration areas, joining the experience of psychologists, neuroscientists, philosophers, and even engineers. The concept of a robot as a cognitive companion for humans may be very useful. The interaction between humans and cognitive robots may be a mediator of movement patterns, learning behaviors from demonstrations, group activities, and social behaviors, as far as higher-order concepts such as symbol manipulation capabilities, words acquisition, and sensorimotor knowledge organization. Moreover there is an occupation to check many theories, such as transferring the knowledge and skills between humans and robots. Although several robotic solutions for children have been proposed the diffusion of aforementioned ideas is still limited. The review summarizes the current *Corresponding author: Tomasz Komendziski, Department of Cognitive Science, Nicolaus Copernicus University, Toru, Poland, E-mail: tkomen@umk.pl; and Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland Emilia Mikolajewska: Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland; and Department of Physiotherapy, Ludwik Rydygier Collegium Medium in Bydgoszcz, Nicolaus Copernicus University, Toru, Poland Dariusz Mikolajewski: Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland; Institute of Mechanics and Applied Computer Sciences, Kazimierz Wielki Universit, Bydgoszcz, Poland; and Department of Informatics, Nicolaus Copernicus University, Toru, Poland Joanna Dreszer and Bibianna Balaj: Department of Cognitive Science, Nicolaus Copernicus University, Toru, Poland; and Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland Introduction Developmental disorders in children are characterized by a delay of developmental skills expected to achieve in a particular age or developmental stage. Cognitive robots can be a promising therapeutic method for the treatment of developmental disorders. Robot-based interactions supporting therapy in such children become more and more popular. Robots are used as therapeutic tools useful in the therapy of the various developmental disorders as mediators of movement patterns, group activities, and social behaviors. There is no doubt that a deeper research is needed in the aforementioned area, especially based on interdisciplinary scientific and therapeutic teams and evidence-based medicine (EBM) paradigm. This article aims to assess the current and future role of cognitive robots in the development and rehabilitation of children with developmental disorders. Skills and limitations Observing the activities of other people to pick up a new movement is a common way of acquiring new skills during the development. This process is supported by an action-perception matching mechanism, applied also in the clinical context, as rehabilitative training based on a combination of action perception and execution (e.g. in children with hemiplegia due to cerebral palsy). 94Komendziski et al.: Cognitive robots in children with developmental disorders Multimodal stimuli related to the movement (e.g. visual and acoustic action-related inputs) may facilitate the perception and meaning of the activity observed [1]. Reaching for various objects, and then grasping and manipulating them, is a very important activity. Children learn how to use their hands and various objects in increasingly sophisticated ways, including the exploration of the environment and group behavior. Hands are also important mediators of social contact: touch with, to signify feelings, and to enrich interhuman communication with gestural expression. The most advanced are cognitive skills such as writing or the playing of musical instruments [2]. The mechanisms that decide how the visual system assesses a size of objects at different distances are still unclear. The hypothesis is two-fold: different mechanisms may be involved for near (reachable) objects and far (unreachable) objects [3]. We know that the accuracy of such perception changes with age and distance. The haptic system can be used to calibrate visual size perception during development, and the aforementioned calibration mechanisms differ in children and in adults [3]. The lack of clear vision in children should affect the haptic orientation discrimination [4]. Finger manipulation and counting activity also play a very important role in the acquisition of numerical skills in children, including building motor-based representations [5]. Twelve-month-old infants rely on information about the certainty of goal selection extrapolating about the action goal of other people [6]. Delayed sensory feedback to a simple motor act was regarded as causing the recalibration of sensory-motor synchronization. Instantaneous feedback appeared to precede the motor act but not in children ages 8­11 years and adults. Precision in the simultaneity task also decreased with age [7]. Direct tactile feedback interacts with the auditory spatial localization system (important in people with the auditory sense of space; e.g. due to vision disorders) [8]. Haptic perception constitutes the ability to extract object features through object-related activities. Haptic precision in children is compromised: it is lower during active exploration compared to passive motion. This situation is probably caused by imprecise predicted sensory feedback exploratory movements (disturbance) not compensated in mind until mid-adolescence [9]. Whole perception is a complex process, including also judgments of almost all quantities (length, duration, number, etc.). The influence of the previous knowledge on perception is present already in young children, suggesting a strong context dependency of the perception [10]. Multisensory tasks require the integration of information gathered from different sensors. The selection of sensor modalities is faster and more rewarding at early childhood than decisions based on joint space (i.e. based on the integration of the available sources of information, multisensory integration). Experience concerning the quality of learning in the joint space increases with age and maximizes in adulthood while learning in modalities is then less accurate [11]. The haptic system dominates size discrimination and vision dominates orientation discrimination in young children [12]. The multisensory integration of spatial information occurs in 8-year-old children. Learning may be associated with action demonstrations accompanied by speech (explanation of demonstration). Distinguishing between path-oriented utterances (emphasized source, trajectory, or goal) and manner-oriented utterances (emphasized medium, velocity, or means of motion) is an essential part of such development [13]. Before the age of 8­10 years, a strong unisensory dominance occurs for size and orientation visual-haptic judgments. Visual-auditory adultlike behavior develops later [14]. Children and elderly can perform spatial processing in a very similar way. They usually do it significantly worse than young adults [15]. Naming creates an association between objects and words. Thus, there is a strong link between early wordlearning and conceptual development [16]. Premature birth has been associated with damage in many regions of the cerebral cortex, critical for both visual attention and magnitudes perception (time, space, and number). Strong impairments were found on time estimation and attentional task, whereas numerical discrimination or mapping tasks remained relatively unimpaired [17]. A study by Bisio et al. supports the literature proposing the mirror neuron system as neural substrate for rehabilitation and opens a debate on the rehabilitative treatments. Afferent inputs from periphery may evoke plasticity in the human motor system [18]. An analysis of motor strategies, applied by children with autism spectrum disorders (ASD) when they learn a new motor pattern, showed significant differences in continuous time on target (CTT), distance from target (DT), and distance from path (DP) measures as well as 2D reconstructions of children's trajectories. Children with ASD showed difficulties in planning of overall actions [19]. Children with ASD engage rather in highly perseverative and inflexible behaviors. They may show difficulties in processing information, learning of activities of daily living (ADLs), and cognitive concepts [20]. Self-generated mobility constitutes the critical element necessary for a proper physical, emotional, cognitive, and social development of children [21]. Thus, disorders concerning self-locomotion may increase the risk of developmental delays. Externally supported mobility (power mobility devices) may prevent this delay, but such form of movement may limit the development of the other Komendziski et al.: Cognitive robots in children with developmental disorders95 skills (gross motor skills, movement planning, distance assessment, etc.). Moreover, many gait-supporting devices do not allow free movement within the environment, supporting natural and motivating learning through play. The influence of the developmental disorders to the way and pace of development is so huge that it may completely change their functional and cognitive abilities and the quality of further life. Every efficient way of the therapy is precious and may constitute true breakthrough. Toys, robots, and intelligent environments Despite recent history, social and interactive skills are necessary in many applications where robots interact and collaborate with other robots or humans (development of a cognitive robot companion or educational/therapeutic tools for children) [22]. We do not know if such seminatural social interaction will be complete (due to, for example, awareness in people that their companion is a robot) but may fulfill the requirements of the basic human-like interaction. Robots may emulate cognitive behavior by reproducing various aspects of human or animal learning and behavior. Self-learning robots proved to be successful in, for example, navigation through unknown terrain [23]. The therapeutic use of various types of toys and robots may significantly help some children with neurological disorders [24] if used temporarily and in a self-controlled way. They are available, low cost, fun, well accepted by infants and children, functional, and frequently antiallergenic. Toys and robots can not be tired, bored, and irritated. Their control system, appearance, and exactness may be shaped and modified according to needs. Toys and robots can also be used as manipulation tools allowing children with disabilities to participate in play or educational and therapeutic activities [25]. Despite their popularity, there is still a few research, especially randomized controlled trials (RCTs). One of the earliest is an instrumented block-box toy commonly used to assess the ability to manipulate objects and insert them into holes [26]. Such aids support learning and social behavior. Requirements are simple but still constitute challenge: robot should process external information, gain the activity-responding multiple conditions, behave accordingly to the social values, provide flexible and interactive communication skills, and be able to cope even with an unknown situation. Such requirements need advanced technological solutions, such as multimodal communication skills, self-organizing, self-emergent functions, and almost semiconscious activities [27]. Robotic toys such as Keepon may increase a cognitive flexibility task performance in children with ASD. These children are more engaged and they enjoy more during interaction with the robot compared to the interaction with the adult [28]. Multimodal interactions play a central role in robot-supported neurocognitive development [29]. The behavior of a cognitive robot may emulate the behavior of a child in tutoring situations, creating new effective feedback strategies in the tutoring spotter system [13]. Cognitive robots can promote attention, communication, and social skills in adolescents with ASD [30]. Children with and without disabilities may use a robot to perform the same play activities. Fulfilling the task may indicate a full understanding of the underlying cognitive skills, not only influence of deficits [25]. The perseverative errors of children may result, in part, from their sociocognitive ability [31]. Human language knowledge built into robots covers individual learning about itself and the environment, social learning, and learning of linguistic capability. The aforementioned capabilities develop in continuous feedback cycles of interactions, including influence of the context, conditions, and requisites. Of course, the development of language and cognitive skills needs to take into consideration the context and changes the experience and competence [32]. There is no doubt that artificial intelligent systems should interact naturally with human users, learning from human instructions during context-aware activities [13]. The learning of the category formation and vocabulary acquisition in robots through active interaction with children is useful for therapeutic games involving naming and corrective feedback actions between the robot and the human user (e.g. child) [16]. The various responsiveness of children to social input may significantly influence the way and pace of learning [33]. Cause is a slower adaptation to a novel dynamic environment due to developmental deficit. Some spatial abnormalities in children with cerebral palsy may be caused not only by the disturbance of motor control signals (generating weakness and spasticity) but also may be derived from inefficient anticipatory control strategy related to the congenital nature of the brain deficit. The solution in selected cases of spatial abnormalities may be robot-generated speed-dependent force field increasing the abilities of children (i.e. children adapt their reaching movements to perturbations induced by a robotic device) [34, 35]. The cognitive developmental robotics paradigm associates number words (or tags) to fingers, joining sensorimotor skills (finger counting) and linguistic skills (using number words), which is the importance of sensorimotor skill learning in robots for the acquisition of 96Komendziski et al.: Cognitive robots in children with developmental disorders abstract knowledge such as numbers [5]. An interaction with the humanoid robot KASPAR equipped with skin patches improved tactile event recognition in children with ASD [36]. Robotic embodied cognitive architecture (ACT-R Embodied) allows for gaze-following and level 1 perspective taking built in an embodied robotic system [37]. The humanoid robot platform iCub supports a collaborative research in cognitive development through autonomous exploration and social interaction [38]. Two- to 4-year-old children with ASD spontaneously interact with the robot Keepon, expressing its attention (directing its gaze) and emotions (pleasure and excitement). The assumption that the simple expressiveness of Keepon helps children understand socially meaningful information, which then activate their intact motivation to share interests and feelings with others, may be true. Thus, the simplicity of the interaction may facilitate the development of social communication in children with ASD [28, 39]. School children, both typically developing and with severe motor disabilities, used the robotic arm in a three-task sequence routine to dig objects from a tub of dry macaroni. The study emphasized increased classroom participation, expressive language, and the interest of children in the robot-related tasks. The important conclusion was needed to increase the number of colors and sounds and/or music associated with the robot [40]. The aforementioned results were better compared to other interventions using toys and computer games [41]. Robot-based gait training/learning is an innovative field of research in neurological rehabilitation, especially in children who were unable to perform any autonomous form of locomotion before. In the study of Smania et al. a child assisted by the actuated aid was successful in moving around in his environment, but the energy cost of gait was significantly higher than normality. Thus, more advanced aids such as robotic exoskeleton should be incorporated [42, 43]. Novel intelligent wheelchairs adapted for users with cognitive disabilities and mobility impairment may constitute another breakthrough in support of mobility in children with severe disabilities [44]. The spectrum of robot-based therapeutic applications increases. It provides increased effectivity regarding motor problems diagnosed even in children with fetal alcohol spectrum disorder (FASD) [45]. Complementary technologies such as virtual reality (VR) are effective [20]. Supporting therapeutic games may facilitate learning, enhance attentional control, allow greater cognitive flexibility, perception, attention, task switching, or mental rotation, and better address developmental disorders [46]. Children with gait disorders may modify their activity to the demands of the scenario (e.g. VR-based), but the level and severity of their cognitive and motor disorders limit the extent of this modification [47]. Artificial environments are more advanced alternatives to therapeutic toys and robots. Such environments use mobile toys/robots [29]. Another alternative are virtual robots used to overcome the limitations of physical robots (cost, reliability, and the need for local technical support) [25]. Discussion and directions of further research Cognitive therapeutic robots are not an ideal cure for everything, but in selected cases they can be more effective than traditional therapeutic toys. We do not even know if the human-like robot is the best solution for therapeutic toys. We are sure that the shape, color, and texture matter, but research has to be more detailed. We should also recognize the potential ethical pitfalls: it is possible that an appropriately designed robot caregiver has the potential to contribute more positively to the development of the child than the adult [24]. We do not know if a successful child-robot interaction influences further child-therapist interaction. Generalizing and extrapolating current experiment results is not quite safe: the number of technological interventions and observed improvements is too low to compare them to long-term results of traditional development in more naturalistic settings [48]. The use of selfcontrolled technology to learn new ADLs and cognitive concepts (e.g. time perception and imagination) still constitutes a huge challenge. There is still a need for specific kinds of technologies to learn specific kinds of skills (as far ADLs such as perception of emotions) [20]. There are too many unknown mechanisms and many novel technologies that need to be tested. Some of them are brain derived, such as liquid models of cognitive processes [49, 50] or fuzzy logic [51, 52]. However, there are many commonly used techniques that may be easily adapted to the cognitive therapeutic robots, such as exoskeletons, brain-computer interfaces, neuroprostheses, or eHealth solutions [53, 54]. Current evidences underscore the need for deeper investigations of motor strategies exploited by children with various developmental disabilities when learning a new motor procedure. The aforementioned outcomes and clinical guidelines are important for clinical decision-making. Computational models derived from machine learning and developmental robotics may provide a framework to address its impact on developmental disorders. The authors investigated the Komendziski et al.: Cognitive robots in children with developmental disorders97 application of therapeutic toys and cognitive robots within leading approaches concerning the neurorehabilitation of children with various developmental disorders, such as the Bobath concept [55]. Proposals of new solutions in this area such as "System supporting perceptual-cognitive development of infants and babies" have been awarded (Gold medals at Concours Lepine, Brussel Innova, INPEX and INTARG in 2015). The incorporation of intelligent toys and cognitive robots into semiautomated diagnostics and therapy can increase rehabilitation outcomes. A deeper research is needed in the aforementioned area, especially based on interdisciplinary scientific and therapeutic teams and EBM paradigm. Conclusions The further development of the therapeutic application of cognitive robotics may open new directions to improve the basic knowledge and therapeutic use of action observation, activity repetition and learning, social interaction, and even group behavior. There is a need for an integrated approach of scientists, engineers, and clinicians and for more comprehensive studies, including research on the long-term home use of cognitive therapeutic tools shaped as toys. Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission. Research funding: This work was conducted as a part of work within a project "NeuroPerCog: development of phonematic hearing and working memory in infants and children" (head: Prof. Wlodzislaw Duch). The project is funded by the Polish National Science Centre (DEC2013/08/W/HS6/00333, Symfonia 1). Employment or leadership: None declared. Honorarium: None declared. Competing interests: The funding organization(s) played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Bio-Algorithms and Med-Systems de Gruyter

Cognitive robots in the development and rehabilitation of children with developmental disorders

Loading next page...
 
/lp/de-gruyter/cognitive-robots-in-the-development-and-rehabilitation-of-children-raWrXWHgRO
Publisher
de Gruyter
Copyright
Copyright © 2016 by the
ISSN
1895-9091
eISSN
1896-530X
DOI
10.1515/bams-2016-0010
Publisher site
See Article on Publisher Site

Abstract

Cognitive robots constitute a highly interdisciplinary approach to the issue of therapy of children with developmental disorders. Cognitive robots become more popular, especially in action and language integration areas, joining the experience of psychologists, neuroscientists, philosophers, and even engineers. The concept of a robot as a cognitive companion for humans may be very useful. The interaction between humans and cognitive robots may be a mediator of movement patterns, learning behaviors from demonstrations, group activities, and social behaviors, as far as higher-order concepts such as symbol manipulation capabilities, words acquisition, and sensorimotor knowledge organization. Moreover there is an occupation to check many theories, such as transferring the knowledge and skills between humans and robots. Although several robotic solutions for children have been proposed the diffusion of aforementioned ideas is still limited. The review summarizes the current *Corresponding author: Tomasz Komendziski, Department of Cognitive Science, Nicolaus Copernicus University, Toru, Poland, E-mail: tkomen@umk.pl; and Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland Emilia Mikolajewska: Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland; and Department of Physiotherapy, Ludwik Rydygier Collegium Medium in Bydgoszcz, Nicolaus Copernicus University, Toru, Poland Dariusz Mikolajewski: Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland; Institute of Mechanics and Applied Computer Sciences, Kazimierz Wielki Universit, Bydgoszcz, Poland; and Department of Informatics, Nicolaus Copernicus University, Toru, Poland Joanna Dreszer and Bibianna Balaj: Department of Cognitive Science, Nicolaus Copernicus University, Toru, Poland; and Neurocognitive Laboratory, Interdisciplinary Center for Modern Technologies, Nicolaus Copernicus University, Toru, Poland Introduction Developmental disorders in children are characterized by a delay of developmental skills expected to achieve in a particular age or developmental stage. Cognitive robots can be a promising therapeutic method for the treatment of developmental disorders. Robot-based interactions supporting therapy in such children become more and more popular. Robots are used as therapeutic tools useful in the therapy of the various developmental disorders as mediators of movement patterns, group activities, and social behaviors. There is no doubt that a deeper research is needed in the aforementioned area, especially based on interdisciplinary scientific and therapeutic teams and evidence-based medicine (EBM) paradigm. This article aims to assess the current and future role of cognitive robots in the development and rehabilitation of children with developmental disorders. Skills and limitations Observing the activities of other people to pick up a new movement is a common way of acquiring new skills during the development. This process is supported by an action-perception matching mechanism, applied also in the clinical context, as rehabilitative training based on a combination of action perception and execution (e.g. in children with hemiplegia due to cerebral palsy). 94Komendziski et al.: Cognitive robots in children with developmental disorders Multimodal stimuli related to the movement (e.g. visual and acoustic action-related inputs) may facilitate the perception and meaning of the activity observed [1]. Reaching for various objects, and then grasping and manipulating them, is a very important activity. Children learn how to use their hands and various objects in increasingly sophisticated ways, including the exploration of the environment and group behavior. Hands are also important mediators of social contact: touch with, to signify feelings, and to enrich interhuman communication with gestural expression. The most advanced are cognitive skills such as writing or the playing of musical instruments [2]. The mechanisms that decide how the visual system assesses a size of objects at different distances are still unclear. The hypothesis is two-fold: different mechanisms may be involved for near (reachable) objects and far (unreachable) objects [3]. We know that the accuracy of such perception changes with age and distance. The haptic system can be used to calibrate visual size perception during development, and the aforementioned calibration mechanisms differ in children and in adults [3]. The lack of clear vision in children should affect the haptic orientation discrimination [4]. Finger manipulation and counting activity also play a very important role in the acquisition of numerical skills in children, including building motor-based representations [5]. Twelve-month-old infants rely on information about the certainty of goal selection extrapolating about the action goal of other people [6]. Delayed sensory feedback to a simple motor act was regarded as causing the recalibration of sensory-motor synchronization. Instantaneous feedback appeared to precede the motor act but not in children ages 8­11 years and adults. Precision in the simultaneity task also decreased with age [7]. Direct tactile feedback interacts with the auditory spatial localization system (important in people with the auditory sense of space; e.g. due to vision disorders) [8]. Haptic perception constitutes the ability to extract object features through object-related activities. Haptic precision in children is compromised: it is lower during active exploration compared to passive motion. This situation is probably caused by imprecise predicted sensory feedback exploratory movements (disturbance) not compensated in mind until mid-adolescence [9]. Whole perception is a complex process, including also judgments of almost all quantities (length, duration, number, etc.). The influence of the previous knowledge on perception is present already in young children, suggesting a strong context dependency of the perception [10]. Multisensory tasks require the integration of information gathered from different sensors. The selection of sensor modalities is faster and more rewarding at early childhood than decisions based on joint space (i.e. based on the integration of the available sources of information, multisensory integration). Experience concerning the quality of learning in the joint space increases with age and maximizes in adulthood while learning in modalities is then less accurate [11]. The haptic system dominates size discrimination and vision dominates orientation discrimination in young children [12]. The multisensory integration of spatial information occurs in 8-year-old children. Learning may be associated with action demonstrations accompanied by speech (explanation of demonstration). Distinguishing between path-oriented utterances (emphasized source, trajectory, or goal) and manner-oriented utterances (emphasized medium, velocity, or means of motion) is an essential part of such development [13]. Before the age of 8­10 years, a strong unisensory dominance occurs for size and orientation visual-haptic judgments. Visual-auditory adultlike behavior develops later [14]. Children and elderly can perform spatial processing in a very similar way. They usually do it significantly worse than young adults [15]. Naming creates an association between objects and words. Thus, there is a strong link between early wordlearning and conceptual development [16]. Premature birth has been associated with damage in many regions of the cerebral cortex, critical for both visual attention and magnitudes perception (time, space, and number). Strong impairments were found on time estimation and attentional task, whereas numerical discrimination or mapping tasks remained relatively unimpaired [17]. A study by Bisio et al. supports the literature proposing the mirror neuron system as neural substrate for rehabilitation and opens a debate on the rehabilitative treatments. Afferent inputs from periphery may evoke plasticity in the human motor system [18]. An analysis of motor strategies, applied by children with autism spectrum disorders (ASD) when they learn a new motor pattern, showed significant differences in continuous time on target (CTT), distance from target (DT), and distance from path (DP) measures as well as 2D reconstructions of children's trajectories. Children with ASD showed difficulties in planning of overall actions [19]. Children with ASD engage rather in highly perseverative and inflexible behaviors. They may show difficulties in processing information, learning of activities of daily living (ADLs), and cognitive concepts [20]. Self-generated mobility constitutes the critical element necessary for a proper physical, emotional, cognitive, and social development of children [21]. Thus, disorders concerning self-locomotion may increase the risk of developmental delays. Externally supported mobility (power mobility devices) may prevent this delay, but such form of movement may limit the development of the other Komendziski et al.: Cognitive robots in children with developmental disorders95 skills (gross motor skills, movement planning, distance assessment, etc.). Moreover, many gait-supporting devices do not allow free movement within the environment, supporting natural and motivating learning through play. The influence of the developmental disorders to the way and pace of development is so huge that it may completely change their functional and cognitive abilities and the quality of further life. Every efficient way of the therapy is precious and may constitute true breakthrough. Toys, robots, and intelligent environments Despite recent history, social and interactive skills are necessary in many applications where robots interact and collaborate with other robots or humans (development of a cognitive robot companion or educational/therapeutic tools for children) [22]. We do not know if such seminatural social interaction will be complete (due to, for example, awareness in people that their companion is a robot) but may fulfill the requirements of the basic human-like interaction. Robots may emulate cognitive behavior by reproducing various aspects of human or animal learning and behavior. Self-learning robots proved to be successful in, for example, navigation through unknown terrain [23]. The therapeutic use of various types of toys and robots may significantly help some children with neurological disorders [24] if used temporarily and in a self-controlled way. They are available, low cost, fun, well accepted by infants and children, functional, and frequently antiallergenic. Toys and robots can not be tired, bored, and irritated. Their control system, appearance, and exactness may be shaped and modified according to needs. Toys and robots can also be used as manipulation tools allowing children with disabilities to participate in play or educational and therapeutic activities [25]. Despite their popularity, there is still a few research, especially randomized controlled trials (RCTs). One of the earliest is an instrumented block-box toy commonly used to assess the ability to manipulate objects and insert them into holes [26]. Such aids support learning and social behavior. Requirements are simple but still constitute challenge: robot should process external information, gain the activity-responding multiple conditions, behave accordingly to the social values, provide flexible and interactive communication skills, and be able to cope even with an unknown situation. Such requirements need advanced technological solutions, such as multimodal communication skills, self-organizing, self-emergent functions, and almost semiconscious activities [27]. Robotic toys such as Keepon may increase a cognitive flexibility task performance in children with ASD. These children are more engaged and they enjoy more during interaction with the robot compared to the interaction with the adult [28]. Multimodal interactions play a central role in robot-supported neurocognitive development [29]. The behavior of a cognitive robot may emulate the behavior of a child in tutoring situations, creating new effective feedback strategies in the tutoring spotter system [13]. Cognitive robots can promote attention, communication, and social skills in adolescents with ASD [30]. Children with and without disabilities may use a robot to perform the same play activities. Fulfilling the task may indicate a full understanding of the underlying cognitive skills, not only influence of deficits [25]. The perseverative errors of children may result, in part, from their sociocognitive ability [31]. Human language knowledge built into robots covers individual learning about itself and the environment, social learning, and learning of linguistic capability. The aforementioned capabilities develop in continuous feedback cycles of interactions, including influence of the context, conditions, and requisites. Of course, the development of language and cognitive skills needs to take into consideration the context and changes the experience and competence [32]. There is no doubt that artificial intelligent systems should interact naturally with human users, learning from human instructions during context-aware activities [13]. The learning of the category formation and vocabulary acquisition in robots through active interaction with children is useful for therapeutic games involving naming and corrective feedback actions between the robot and the human user (e.g. child) [16]. The various responsiveness of children to social input may significantly influence the way and pace of learning [33]. Cause is a slower adaptation to a novel dynamic environment due to developmental deficit. Some spatial abnormalities in children with cerebral palsy may be caused not only by the disturbance of motor control signals (generating weakness and spasticity) but also may be derived from inefficient anticipatory control strategy related to the congenital nature of the brain deficit. The solution in selected cases of spatial abnormalities may be robot-generated speed-dependent force field increasing the abilities of children (i.e. children adapt their reaching movements to perturbations induced by a robotic device) [34, 35]. The cognitive developmental robotics paradigm associates number words (or tags) to fingers, joining sensorimotor skills (finger counting) and linguistic skills (using number words), which is the importance of sensorimotor skill learning in robots for the acquisition of 96Komendziski et al.: Cognitive robots in children with developmental disorders abstract knowledge such as numbers [5]. An interaction with the humanoid robot KASPAR equipped with skin patches improved tactile event recognition in children with ASD [36]. Robotic embodied cognitive architecture (ACT-R Embodied) allows for gaze-following and level 1 perspective taking built in an embodied robotic system [37]. The humanoid robot platform iCub supports a collaborative research in cognitive development through autonomous exploration and social interaction [38]. Two- to 4-year-old children with ASD spontaneously interact with the robot Keepon, expressing its attention (directing its gaze) and emotions (pleasure and excitement). The assumption that the simple expressiveness of Keepon helps children understand socially meaningful information, which then activate their intact motivation to share interests and feelings with others, may be true. Thus, the simplicity of the interaction may facilitate the development of social communication in children with ASD [28, 39]. School children, both typically developing and with severe motor disabilities, used the robotic arm in a three-task sequence routine to dig objects from a tub of dry macaroni. The study emphasized increased classroom participation, expressive language, and the interest of children in the robot-related tasks. The important conclusion was needed to increase the number of colors and sounds and/or music associated with the robot [40]. The aforementioned results were better compared to other interventions using toys and computer games [41]. Robot-based gait training/learning is an innovative field of research in neurological rehabilitation, especially in children who were unable to perform any autonomous form of locomotion before. In the study of Smania et al. a child assisted by the actuated aid was successful in moving around in his environment, but the energy cost of gait was significantly higher than normality. Thus, more advanced aids such as robotic exoskeleton should be incorporated [42, 43]. Novel intelligent wheelchairs adapted for users with cognitive disabilities and mobility impairment may constitute another breakthrough in support of mobility in children with severe disabilities [44]. The spectrum of robot-based therapeutic applications increases. It provides increased effectivity regarding motor problems diagnosed even in children with fetal alcohol spectrum disorder (FASD) [45]. Complementary technologies such as virtual reality (VR) are effective [20]. Supporting therapeutic games may facilitate learning, enhance attentional control, allow greater cognitive flexibility, perception, attention, task switching, or mental rotation, and better address developmental disorders [46]. Children with gait disorders may modify their activity to the demands of the scenario (e.g. VR-based), but the level and severity of their cognitive and motor disorders limit the extent of this modification [47]. Artificial environments are more advanced alternatives to therapeutic toys and robots. Such environments use mobile toys/robots [29]. Another alternative are virtual robots used to overcome the limitations of physical robots (cost, reliability, and the need for local technical support) [25]. Discussion and directions of further research Cognitive therapeutic robots are not an ideal cure for everything, but in selected cases they can be more effective than traditional therapeutic toys. We do not even know if the human-like robot is the best solution for therapeutic toys. We are sure that the shape, color, and texture matter, but research has to be more detailed. We should also recognize the potential ethical pitfalls: it is possible that an appropriately designed robot caregiver has the potential to contribute more positively to the development of the child than the adult [24]. We do not know if a successful child-robot interaction influences further child-therapist interaction. Generalizing and extrapolating current experiment results is not quite safe: the number of technological interventions and observed improvements is too low to compare them to long-term results of traditional development in more naturalistic settings [48]. The use of selfcontrolled technology to learn new ADLs and cognitive concepts (e.g. time perception and imagination) still constitutes a huge challenge. There is still a need for specific kinds of technologies to learn specific kinds of skills (as far ADLs such as perception of emotions) [20]. There are too many unknown mechanisms and many novel technologies that need to be tested. Some of them are brain derived, such as liquid models of cognitive processes [49, 50] or fuzzy logic [51, 52]. However, there are many commonly used techniques that may be easily adapted to the cognitive therapeutic robots, such as exoskeletons, brain-computer interfaces, neuroprostheses, or eHealth solutions [53, 54]. Current evidences underscore the need for deeper investigations of motor strategies exploited by children with various developmental disabilities when learning a new motor procedure. The aforementioned outcomes and clinical guidelines are important for clinical decision-making. Computational models derived from machine learning and developmental robotics may provide a framework to address its impact on developmental disorders. The authors investigated the Komendziski et al.: Cognitive robots in children with developmental disorders97 application of therapeutic toys and cognitive robots within leading approaches concerning the neurorehabilitation of children with various developmental disorders, such as the Bobath concept [55]. Proposals of new solutions in this area such as "System supporting perceptual-cognitive development of infants and babies" have been awarded (Gold medals at Concours Lepine, Brussel Innova, INPEX and INTARG in 2015). The incorporation of intelligent toys and cognitive robots into semiautomated diagnostics and therapy can increase rehabilitation outcomes. A deeper research is needed in the aforementioned area, especially based on interdisciplinary scientific and therapeutic teams and EBM paradigm. Conclusions The further development of the therapeutic application of cognitive robotics may open new directions to improve the basic knowledge and therapeutic use of action observation, activity repetition and learning, social interaction, and even group behavior. There is a need for an integrated approach of scientists, engineers, and clinicians and for more comprehensive studies, including research on the long-term home use of cognitive therapeutic tools shaped as toys. Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission. Research funding: This work was conducted as a part of work within a project "NeuroPerCog: development of phonematic hearing and working memory in infants and children" (head: Prof. Wlodzislaw Duch). The project is funded by the Polish National Science Centre (DEC2013/08/W/HS6/00333, Symfonia 1). Employment or leadership: None declared. Honorarium: None declared. Competing interests: The funding organization(s) played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication.

Journal

Bio-Algorithms and Med-Systemsde Gruyter

Published: Sep 1, 2016

References