Introduction

The human body focuses scientific attention in many disciplines and in interdisciplinary research. As it is the core and material foundation of human existence, it has also become the subject of numerous studies in the field of psychology and neuropsychology (or in broadly conceived neuroscience). Cognitive approach to body experience emphasizes both its complexity and specificity. First, one’s body experience relies on multisensory data integration (Gallagher and Cole 1995; Jeannerod 2004; Holmes and Spence 2004; Giummarra et al. 2008; Medina and Coslett 2010) which result in coherent physical self and create sense of body ownership (Haggard et al. 2003; Botvinick 2004; Press et al. 2008; Maselli 2015). Second, body-related data processing involves networks of many brain regions which are spread in the human brain (Churchland 2002). Among them fronto-parietal circuit was evidenced to play predominant role (Rizzolatti et al. 1997; Jeannerod 1999; Galati et al. 2001; Mohr et al. 2006; Tsakiris et al. 2007; Kemmerer and Tranel 2008). Third, human body (faces and other body parts) gains a special status as socially relevant stimuli which play a privileged role in perceptual processing (Bracco and Chiorri 2009; Boyer et al. 2017). As a result, body is a specific mental category within which data are processed faster comparing to other objects of a similar complexity (Shontz and McNish 1972; Ro et al. 2007). Therefore, human minds create and store distinct representations for body and other physical objects (Semenza and Goodglass 1985; Reed and Farah 1995; Coslett 1998; Corradi-Dell’Acqua and Rumiati 2007).

Despite long history and tradition of body representation in psychological and neurological literature (Semenza and Delazer 2003; Vallar and Papagno 2003; Holmes and Spence 2006; Maravita 2006; Gallagher 2006; Corradi-Dell’Acqua and Rumiati 2007; Longo et al. 2010), there is still a lively discussion about its specificity and characteristics. Although results of studies provide evidence that body representation should not be considered as a homogeneous concept but rather as a multidimensional and complex one, with various types to be distinguished, there is still much confusion about these types. Firstly, researchers use many terms and definitions to discuss the types of body representation, and it is not clear to what extent these types are distinct from one another. Such ambiguities result in the usage of manifold body representation measures, which makes the results of studies difficult to compare. Secondly, relations between types of body representation need to be carefully investigated. However, it is evidenced that types of body representation may be dissociated as a consequence of brain injuries (Sirigu et al. 1991; Reed and Farah 1995; Denes et al. 2000; Guariglia et al. 2002; Schwoebel and Coslett 2005; Corradi-Dell’Acqua and Rumiati 2007; Longo et al. 2010); some patients may also reveal deficits in more than one body representation type (Schwoebel and Coslett 2005).

The study by Schwoebel and Coslett (2005) is an attempt to resolve some of the above doubts by investigating types of body representation in patients after brain injuries. Based on recent evidence, the authors assume that there are three types of body representation, which differ in their contents: body schema, body structural description, and body semantics. Body schema is described as a dynamic representation of body part positions, which is derived from sensory data (tactile, visual, vestibular, and proprioceptive) and contributes to the motor system (Schwoebel and Coslett 2005). It gives a basis for the performance of movements by providing information about the dynamic position of each body part in relation to other parts (Sirigu et al. 1991; Denes et al. 2000; Schwoebel and Coslett 2005; Laiacona et al. 2006). As a result, it plays crucial role for body ownership, i.e. the sense that one’s body belongs to oneself (see, for instance, van Stralen et al. 2013; Burin et al. 2015; Llorens et al. 2017). It is created in real time and serves as a basis also for movement imagining (Coslett 1998; Schwoebel et al. 2004; Schwoebel and Coslett 2005; Laiacona et al. 2006). Body schema, similarly to body image, is one of the terms which is the longest used in a scientific reflection on mental models of body. Nonetheless, body schema is diversely defined in philosophical, neurological, and psychological literature and varied relationships between body schema and body image are described. It is noteworthy that some definitions of body schema often do not include postural and dynamic component, but rather emphasize perception of body size and configuration of body parts (Brytek-Matera and Rybicka-Klimczyk 2009) or an ability to experience body boundaries and to locate a stimuli on the surface of one’s own body (see Wolak 1989). In the present study, body schema is understood, following neuropsychological approach, as online and dynamic representation derived from multisensory data which is clearly distinct from map of the body, i.e. body structural description. The latter contains data (mainly visual) about the location of body parts and defines the borders of the body as well as distance relations between body parts (Buxbaum and Coslett 2001; Schwoebel et al. 2004). Body semantics comprises the names of body parts and verbally coded knowledge about body parts (their functions, relations, and associations with objects). Some authors suggest that body semantics also contains information about the typical shape and appearance of body parts (Kemmerer and Tranel 2008). These types of body representation are not only based on various data and processes but also seem to have specific and distinct brain activity patterns. As suggested by Schwoebel’s and Coslett’s study (2005), body image and body structural description are lateralized to the left hemisphere, whereas body schema in not clearly lateralized. Moreover, as the research indicated, body schema deficits were associated with lesions of parietal lobes and/or dorsolateral frontal cortex, whereas form and lexical–semantic knowledge of the body was impaired mainly by temporal lesions (Schwoebel and Coslett 2005).

Systematic group studies on body representation in individuals with brain injuries, such as the research conducted and described by Schwoebel and Coslett (2005), are still rare. Recently, the effects of rubber hand illusion in stroke patients were described (Burin et al. 2015; Llorens et al. 2017), but these reports focus exclusively on reconfiguration of body schema and changes in body ownership. Body representation deficits are analysed mainly in single case studies (Semenza 1988; Chatterjee 1996; Mozaz and Morris 1997; Denes et al. 2000; Buxbaum and Coslett 2001; Guariglia et al. 2002; Felician et al. 2003; Laiacona et al. 2006), which restricts the possibility to drawing more general conclusions on the organization of body representation. There are also doubts connected with the number and content of body representation types as well as with the validity of body representation measures, especially in the case of other concurrent neuropsychological dysfunctions. For this reason and due to the ambiguity of previous conclusions, the current study was undertaken with the following objectives: (1) to describe the consequences of vascular brain injury for body representation and to estimate the prevalence of body representation deficits in patients after strokes; (2) to provide evidence for different types of body representation; (3) to identify the relationship between body representation disorders and other neuropsychological dysfunctions. To accomplish the above aims, 50 patients after strokes and fifty healthy controls were examined with a battery of tasks designed to assess body schema, body structural description and body semantics. As described in “Appendix”, tasks to investigate body schema included hand action and hand laterality tasks. They required motor performance with both hands and motor imagery, respectively. Tasks developed to investigate body structural description involved body parts localization and tactile input localization, both on own body and on the body in a photograph, and matching body parts by location. To assess body semantics, tasks requiring subjects to use semantic and verbally coded body knowledge were applied. They included naming and designating body parts, and matching body parts to objects. In line with the previous literature, it was assumed that brain injury affects body representation and results in deficits of body-related data processing. Since information about the percentage of patients after vascular brain injuries which are impaired on body representation is scarce, a question about frequency of body representation deficits was also posed. As Schwoebel and Coslett (2005) suggest, even more than a half of stroke patients may be affected in at least one body representation measure. In the current study, comparing the scores of patients with these of healthy individuals, the number of stroke patients with body representation deficits was assessed. Further, although the existence of functionally distinct body representation types is suggested in the literature, there is still short evidence considering kinds of body representation components which may be distinguished. To investigate relationships between body representation measures and, as a result, to explore putative types of body representation composed with distinct body-related data, a principal component analysis was performed. Moreover, there have been no clear conclusions considering relationships between body representation deficits and other cognitive dysfunctions. Some of the reports suggest that the former may co-occur with the latter and reflect a more global deficit (see Reed and Farah 1995). On the other hand, some case studies describe patients impaired on body representation measures but intact in other cognitive domains (e.g. Sirigu et al. 1991).

Summing up, it was aimed to make at least three essential contributions with the current study. First, the research is a response to the need for further systematic exploration on the impact of vascular brain injuries on body representation. This is important, as it has been noted, that reports on how often and to what extent strokes impair body representation are not sufficiently discussed in the literature. Second, the research builds on the knowledge on types of body representation, their content, and specificity. It was aimed to investigate structure of body representation in brain-injured individuals with a wide set of tasks considering diversity of body-related data. Third, as there have been little group studies which provide evidence on relationships between body representation deficits and other cognitive dysfunctions, the present study responds to a call for further research in this field. Whereas most prior research has reported results from single case studies, the current one makes an attempt to build on the literature by providing results of a group analysis which sheds light on the deficits co-occurrence.

Method

Participants

Two groups participated in the study: patients after vascular brain injuries and healthy controls. All of them were Polish and White.

Patients with brain injuries

Because of the specificity of the study, purposive sampling was applied using two main including criteria: vascular brain injury and the patients’ general condition enabling them to sit at the table during the assessment for at least 20 min. There were a number of excluding criteria, namely left-handedness, uncorrected ocular defects or visual deficits, epilepsy, chronic alcohol abuse, cognitive deterioration, dementia, neurodegenerative diseases, past traumatic brain injuries with neurological consequences, and a neuropsychological state that made the examination impossible (e.g. acute confusion, agitation, severe aphasia, or memory impairment). Eventually, the group consisted of 50 participants (24 women and 26 men). Their age ranged from 43 to 85 (M = 67.71, SD = 9.65). The mean time that had elapsed from a stroke was 5.35 months (from 1 to 56 months, SD = 10.35), and the mean time spent in the rehabilitation was 22.4 days (from 1 to 73 days, SD = 18.9). The characteristics of the group, based on medical documentation, are presented in Table 1.

Table 1 The characteristic of the group of patients after vascular brain injuries (n = 50)
Full size table

The mean level of muscle tension in the Lovett’s scale (Cuthbert and Goodheart 2007; Zimmerman-Górska 2012) was 3.33 in the upper limbs (min = 0, max = 5, SD = 1.48) and 3.74 in the lower limbs (min = 1.5; max = 5, SD = 0.93). None of the patients reported body experience dysfunctions, such as somatoparaphrenia, anosognosia, limb personification, kinaesthetic hallucinations, or supernumerary phantom limbs.Footnote 1 One person reported an aversion to a disabled limb (which may be treated as misoplegia). Most of the patients complained about the oddity of an unfit limb, which was sometimes described as “weaker and thicker”, “different than the other one”, “ill at ease”, “dead”, “a pickaxe”, “like not mine”, or “not listening to me”. Most of the patients suffered from other health problems, mainly from cardiovascular diseases (98% of participants), endocrine diseases (58%), and locomotor system diseases (osteoarthritis—24% and osteoporosis—4%).

Controls

The group consisted of 50 healthy right-handed adults (27 women and 23 men). The following excluding criteria were applied: cognitive deterioration, dementia, neurodegenerative diseases, past traumatic brain injuries with neurological consequences, stroke, epilepsy, locomotor system injuries and diseases that caused disabilities, not corrected ocular defects or visual deficits, and chronic alcohol abuse. The mean age was 68.17 years (from 43.50 to 85.83, SD = 11.57). A part of this group suffered from cardiovascular diseases (76%), endocrine diseases (22%), and locomotor system diseases (38% had osteoarthritis).

The two groups did not differ significantly in terms of mean age and sex distribution. Patients after vascular brain injuries who participated in the study did not differ from healthy controls in designating objects and in visuo-spatial functions, but they exhibited a significantly lower level of attention, had longer reaction times, and scored lower on naming objects, executing commands, and working memory tasks. All measures used for general neuropsychological assessment are described further in “Measures” section and in "Appendix". For detailed scores and their comparisons between the groups, please see Table 6 in "Appendix".

Measures

The methods were divided into two groups: (1) methods of general neuropsychological assessment and (2) methods of body representation examination categorized in the three dimensions of body representation: body schema, body structural description, and body semantics.

Methods of general neuropsychological assessment

The tasks had a paper-and-pencil as well as electronic form. The set included tasks assessing the following cognitive functions: attention and simple motor reactions, language, working memory, and visuo-spatial functions. All tasks, except for the measure of working memory, were designed by author for the present study. The tasks are described in details in “Appendix”.

Methods of body representation examination

The methods of body representation examination were divided into three groups according to the types of body representation (body schema, body structural description, and body semantics). They were designed by the author for the present study and were patterned upon the tasks used by Schwoebel and Coslett (2005). Some modifications were made, both in the number and in the structure of the tasks. All tasks were constructed in order to minimize the usage of verbal communication (when giving the answers) and have a paper-and-pencil or computer form. The stimuli for the body representation tasks were life-like colourful photographs of real body parts and objects displayed on the computer screen and a printed colourful photograph of a human body (A4 size) exposed on the table in front of the subjects. All tasks are elaborated in “Appendix”.

Procedure

The study had an ex post facto design. The participants were selected by purposive sampling, based on including and excluding criteria (see Participants). All of them gave informed consent. Participation in the study was voluntary, and the respondents did not receive any reward. Before tasks of neuropsychological assessment and body representation were applied, structured interview, observation, and medical data survey were used to receive basic information about the participants and to decide about including them in the study. The interview provided information about subjects’ age, education, handedness, health problems, visual/hearing impairments, self-reliance, previous brain injuries, and drug use. It was also aimed at assessing the participant’s orientation. In the group of patients with brain injuries, questions referring to body experience were also asked. The second part of the interview with participants older than 55 in both groups included the questions of Reisberg’s Global Deterioration Scale (Reisberg et al. 1988; Barcikowska and Bilikiewicz 2004). The observation sheet was used only in the group of patients with vascular brain injuries and included categories such as motor impairments, activity, making contact, communication, and verbal behaviours, following the instructions, reactions and emotional expression, health complaints, untypical behaviours, and other difficulties. In the group of patients, medical data survey was applied. The data were based on medical documentation and considered type and date of the vascular incident, duration of rehabilitation, result of neuroimaging examination, previous brain injuries and their consequences, the patient’s neurological state, self-reliance, and other health problems. The group of participants after vascular brain injuries consisted of patients of the Upper Silesian Rehabilitation Centre in Tarnowskie Góry. They were examined individually, during one, two, or three sessions, depending on a patient’s condition and fatigue. One session lasted an average of 35 min. The assessment of healthy controls was carried out during one meeting that lasted from 40 to 60 min. The measures were ordered in such a way that the tasks which were similar and used the same materials did not occur consecutively.

The study was approved by the Ethics Committee of the Faculty of Pedagogy and Psychology of Maria Curie-Sklodowska University in Lublin.

Statistical analyses

The following statistics were applied to explore and compare the scores: frequencies, descriptive statistics, the Kolmogorov–Smirnov test, and the independent t test and Mann–Whitney U test for statistical significance. To examine the structure of body representation in the group of patients after vascular brain injuries, a principal component analysis was applied. Data of 47 participants were taken into account, as three patients were excluded due to missing data. For eight body representation tasks, ten measures were considered, because in two tasks (localizing body parts and tactile input localization) pointing to one’s own body and pointing to the body parts in a photograph were analysed independently. For nine tasks (hand laterality, localizing body parts, tactile input localization, matching body parts by location, naming body parts, designating body parts, and matching body parts to objects), the numbers of correct answers were treated as measures. In hand action task, the number of incorrect movements was computed. For all tasks, standardized results were analysed. To explore the relationships between body representation and cognitive functioning in patients after brain injuries, the co-occurrence of body representation deficits and other neuropsychological deficits with Chi-square test was analysed. It was assumed that a deficit occurred when a patient’s score fell two standard deviations below the mean score of healthy subjects. Additionally, correlations between the results in general neuropsychological assessment tasks and body representation tasks as well as between time since injury and body representation measures were calculated with Pearson and Spearman correlation coefficients.

Data were computed with IBM SPSS Statistics 20.

Results

Body representation: intergroup comparisons

Patients after vascular brain injures differed significantly from healthy controls in total scores of all body representation measures. The participants’ results and the comparisons are summarized in Table 2.

Table 2 Mean scores of stroke patients and healthy controls in body representation measures
Full size table

Body representation in patients after brain injuries: intragroup analyses

Time since injury and body representation

Median split on time since injury was used to divide subjects into two groups (Me = 2.01 months) to explore whether the time may affect results in body representation measures. Comparisons revealed that there were no statistically significant differences in body representation tasks between patients who experienced stroke less than 2 months and more than 2 months before the assessment. Furthermore, in the sample there were no significant correlations between time since injury and body representation scores.

Principal component analysis

The appropriateness of component analysis for the observed correlations was suggested by the value of .70 resulting from the Keiser–Meyer–Olkin measure of sampling adequacy. Examination suggested three components, which accounted for 66.55% of the variance in the results. Component loadings are presented in Table 3.

Table 3 Component loadings (C1–C3) and the percentage of variance accounted for principal component analysis with varimax rotation
Full size table

The results of principal component analysis and their possible explanations will be discussed further in “Discussion” section.

Body representation deficits

Subsequently, it was checked whether there were deficits in single body representation tasks in patients after brain injuries. Performance of the patients was compared to the performance of healthy individuals. The scores that were at least two standard deviations below the mean score of healthy controls were considered as an indicator of a deficit in the task. Interestingly, only nine out of 47 patients (19%) were not impaired in any of the ten body representation measures. Data about deficits in body representation measures are presented in Table 4.

Table 4 Number and percentage of patients with deficits in body representation measures (n = 47)
Full size table

Relationships between body representation deficits and neuropsychological variables

Ten patients had neuropsychological deficits without body representation deficits. In four participants after brain injuries, the pattern of deficits was opposite: there were a deficit in at least one body representation measure but no other neuropsychological deficits. Analysis with Chi-square test revealed that there existed statistically significant relationships between some deficits. Chi-square test values for these relationships are presented in Table 5.

Table 5 Chi-square values for the significant relationships between neuropsychological deficits and body representation deficits
Full size table

For these statistically significant relationships, correlation coefficients between total scores in the tasks were computed. As analyses showed, only two correlations were statistically significant: between scores in the visuo-spatial functions task and in the matching body parts by location task (ρ = .64, p < .01) and between scores in the visuo-spatial functions task and the task of matching body parts to objects (ρ = .48, p < .01).

Discussion

The main aim of the study was to describe the state and structure of body representation in patients after vascular brain injuries. The results obtained by patients and healthy participants in tasks designed to measure various types of body representation were computed and analysed. The study was based on the work of Schwoebel and Coslett (2005), which, to the best of my knowledge, is the only study on various types of body representation in a large group of participants after brain injuries. It was planned to test the model that distinguishes three types of body representation: body schema, body structural description, and body semantics, as well as to compare the results to those obtained by Schwoebel and Coslett in their study.

Firstly, intergroup comparisons revealed that vascular brain injury results in body representation disintegration, which is manifested in deficits of: (1) movement planning, movement performance, and movement control for both hands; (2) imagining the movements of one’s body and performing its mental rotation; (3) the ability to localize body parts and to apprehend spatial relationships between them; (4) designating and naming body parts, as well as deficits in semantic knowledge about body parts, including the associations of body parts with objects. Because there is still a scarcity of studies on body representation in patients after brain injuries, and because epidemiological data on body representation deficits in this group are still rare and remain inconsistent, the question about the frequency of body representation deficits was valid. Deficit in at least one body representation task appeared in 81% of patients after bran injuries. The rate is much higher than the 51% found in Schwoebel and Coslett’s study (2005). It can therefore be concluded that some kind of body representation deficit is a common consequence of brain injury. In the present study, this referred the most often to dynamic body representation, which enables a person to perform previously planned movements. This deficit manifests itself in difficulties in moving one’s hand on command (in 60% of patients). The result remains in line with the research of Llorens et al. (2017), which indicates that stroke is a common cause of body schema disorder, influenced by motor impairment and a suppression of the reflex activity. In more than a half of the patients, deficits involved naming body parts and matching body parts by location. The least frequently observed deficits (in 25% of patients) concerned tactile input localization, designating body parts, and determining the side of a hand (in hand laterality task). Data showing the infrequent prevalence of impairments in understanding terms describing body parts are consistent with previous reports. It was proven that some aspects of body representation (e.g. designating body parts) are more resistant to distortions resulting from brain injury (Kemmerer and Tranel 2008). There were no correlations between time since injury and body representation state.

Secondly, the results of principal component analysis suggest that it is possible to speak about functionally distinct types or components of body representation, which is in line with the results of previous studies of patients after brain injuries (both group and single case studies) (Sirigu et al. 1991; Coslett 1998; Buxbaum and Coslett 2001; Schwoebel and Coslett 2005). However, the results of the analysis do not confirm the components that were obtained by Schwoebel and Coslett (2005) in 70 stroke patients. Principal component analysis suggests that three components can be distinguished. The first one is loaded by tasks which, according to Schwoebel’s and Coslett’s assumptions, are elements of both body structural description (localizing body parts and matching body parts by location) and body semantics (designating body parts, naming body parts, and matching body parts to objects). When trying to explain and interpret the first factor, one should note that all tasks which load on this component can be characterised by a significant share of the semantic factor. Tasks connected with body semantics require the use of body parts names and the use of knowledge about the associations between body parts and objects (clothes or accessories). Tasks which are usually treated as body structural description measures are also based on concepts and their categorization, as they require the use of semantic elements of body representation, such as verbally coded knowledge about the appearance, shapes, and locations of body parts (Kemmerer and Tranel 2008). Moreover, in the task of matching body parts by location, a participant probably categorizes body parts in terms of body region first (eliminating parts from a different body area) and only later analyses the spatial relationships between body parts. Similarly, pointing to body parts on one’s own body and on a body in a picture presumably requires the use of body part concepts. Furthermore, when explaining the first component, it is important to remark that all tasks which load on this component use specific material, namely photographs of isolated body parts. Such material requires references to single parts only, without the need to analyse the body as a whole. What is more, isolated body parts are neutral elements of general body representation that are not directly connected with the self. It can therefore be stated that tasks which use pictures of isolated body parts and load on the first factor involve the processing of data about the body on a kind of “impersonal” level. The second component is loaded by the tactile input localization task (a measure of body structural description in Schwoebel’s and Coslett’s study) and the hand laterality task (a measure of body schema according to these authors). The first task consists in feeling touch and mapping the input onto body surface, and the latter requires feeling one’s own-body position and performing mental rotations of body parts. It can be concluded that both tasks involve the representation of one’s own body as well as basic body experience. The third component is loaded by two tasks. One is treated as a measure of body schema (the hand action task) and the other as a measure of body structural description (tactile input localization in a photograph). It is noteworthy that the two tasks involve both the representation of one’s own body (movement performance and localization of touch) and a mental model of another person’s body (the investigator’s body or a body presented in a picture). Therefore, they both require a “translation” or “conversion” between “body-as-an-object representation” and “own-body representation” (observed movement vs. one’s own motor programme in the hand action task, and felt touch vs. input localization on another person’s body in the tactile input localization task).

The factor structure of body representation in patients after vascular brain injuries in the present research diverges from the structure described by Schwoebel and Coslett (2005). The difficulty in comparing the results of both studies directly may be a consequence of methodological solutions, e.g. different tasks which were used in the studies. Furthermore, the sample of patients in the current research differed from that of Schwoebel and Coslett (2005) in terms of number of participants, their age, and brain injury location. In the current project, the sample was smaller, the average age of participants was higher, and location of post-stroke injuries was not confined to a single hemisphere which may have contributed to some discrepancies in the results. On the other hand, the findings may initially indicate that components can be marked off not only on the basis of their contents, which is connected with processes involved (motor, visuo-spatial, lexical, and semantic), but presumably also on the basis of their relationship with the self. Body representation seems to consist of various components or elements (or types), which are “saturated” with the self to various degrees. Interpreting the obtained components, I suggest that the first component is the least strongly connected with the self and relates primarily to body as an object, which is processed on an “impersonal” level. This component is associated with “Me” as an object of experience and reflective awareness (Tagini and Raffone 2010). The relationship with the self in the second component is the strongest—the tasks require the processing of sensations that come from one’s own body and are connected with the basic aspects of body experience. Their performance demands updating of the relative positions of one’s body parts (a kind of body map) using both static and dynamic proprioceptive signals, and knowledge of the position of one’s body to programme and guide direct actions towards the body (Paillard 1999; Anema et al. 2009). The component is strongly related to “I” or subjective sense of the self where the body itself is a subject of experience and the first-person perspective is taken (Tagini and Raffone 2010). The third component relates both to the body as part of the self and to representations of other bodies (body as an element of physical and social environment). It comprises the processes of comparing the self with others and conversion between the two aspects of the mental model of a body.

Apart from the abovementioned differences, there is also a similarity between the factor structure of body representation in Schwoebel’s and Coslett’s study and the one described in this text. In both studies, the hand laterality task and the hand action task, which are assumed to measure body schema, loaded on two separate factors (in Schwoebel’s and Coslett’s study each of the mentioned tasks loaded on a distinct factor, which was not loaded by any other task—that is, the factors were loaded by one task only; in my study, the tasks were included in diverse factors, which were also loaded by other tasks). This kind of separateness of these two tasks is inconsistent with the findings reported in the literature, which indicate that both tasks refer in a similar way to dynamic body representation (Parsons 1987; Stephan et al. 1995; Coslett 1998; Jeannerod 1999; Gerardin et al. 2000; Schwoebel et al. 2001, 2002; Amorim et al. 2006; Sekiyama 2006; Creem-Regehr et al. 2007). This issue needs to be explored both in clinical samples and in healthy participants.

Summarizing the results concerning the relationships between body representation and general neuropsychological functioning, it is legitimate to conclude that in a considerable proportion of participants with vascular brain injuries deficits in body representation tasks co-occurred with neuropsychological deficits but only several relationships between the two types of deficits appeared to be statistically significant. There were associations between three body representation tasks (measuring tactile input localization on one’s own body, designating body parts, and determining the side of the hand) and working memory deficit. Furthermore, visuo-spatial deficit was related to performance in the tasks of matching body parts by location, matching body parts to objects, and localizing body parts. Finally, deficit in naming objects was significantly related to deficit in naming body parts. It should be emphasized that in the group of patients with deficits there were only two significant correlations between scores in body representation tasks and general neuropsychological assessment tasks (it referred to visuo-spatial processing and matching body parts by location, as well as visuo-spatial processing and matching body parts to objects). The results suggest that only for the abovementioned pairs one can conclude on the correlations between the level of body representation deficits and cognitive dysfunctions. The extent of distortions in other aspects of body representation does not have to be related to the intensity of concurrent neuropsychological dysfunctions. On the other hand, there were also some participants after brain injuries who exhibited cognitive dysfunctions without body representation distortions, as well as those in whom, conversely, body representation deficits were observed without co-occurring neuropsychological dysfunctions. There were no associations between deficits in attention, executing commands, and designating objects and deficits in body representation tasks. The results of the present study suggest that body representation deficits may co-occur with cognitive deficits, which is in line with the results of previous research (Semenza and Goodglass 1985; Reed and Farah 1995; Denes et al. 2000; Laiacona et al. 2006; Corradi-Dell’Acqua and Rumiati 2007; Kemmerer and Tranel 2008), with the reservation that the latter are a background or possible risk factor for body representation deficits rather than sufficient or imperative requisites for body representation deficits. Moreover, the extent of body representation impairment does not have to be related to the extent of concurrent neuropsychological dysfunctions. Such conclusions are consistent with the observations concerning patients reported in the literature and showing that impairments of body representation with intact functioning of other cognitive domains, as well as reverse situations, are possible (Semenza and Goodglass 1985; Semenza 1988; Suzuki et al. 1997; Guariglia et al. 2002).

To sum up, results of the present study contribute to expand the knowledge about how body representation is affected by vascular brain injuries. It was evidenced that strokes often result in body representation disintegration, which is observed in deficits in various body representation measures. These deficits may be related to other cognitive dysfunctions, but may be also dissociated and independent from them. Moreover, body representation revealed in patients after vascular brain injuries its multifaceted structure, within which distinct types of body representation may be distinguished. The conclusions discussed, particularly the interpretations of body representation components which might be diversely related to the self, should be treated as suggestions which may outline directions for future research. The proposal should be verified and the relationships between body representation and the nature and functioning of self should be explored. Addressing these questions in future studies may provide some advances in the field of associations between various facets of consciousness and self-related processes which still remain uncertain (Tagini and Raffone 2010).

Despite the contribution the presented study makes to the knowledge on body representation, it has its limitations. First, the research groups were not very large. It would be interesting to verify the results in more numerous samples. Second, the group of patients after vascular brain injuries was diverse in terms of age, time which passed from the injury, and brain injury locations. Future studies including more homogenous samples would be beneficial for clarifying the question of body representation structure. Third, methods with unknown validity were used mostly in the study due to specific aims of the research. It would be advisable to investigate psychometric properties of the methods and verify them in further studies. Last, the heterogeneity of the sample in terms of brain injury location and limited information about the location in some cases made it impossible to draw the conclusions on probable brain correlates of the distinguished body representation types.