Aframeworkforexpertsystemknowledgebaseinartpsychotherapy

Chapter 7A Framework for Expert System Knowledge Base in Art PsychotherapySeong-in Kima, Seok Yoob, Kee-Eung Kim c and Youngho Leeaa Division of Information Management Engineering, Korea University 5-1 Anam-dong, Seongbuk-ku, Seoul 136-701, Republic of Korea Tel: +82-18-223-7777; fax: +82-2-428-7772: E-mail address: tenniskorea.ac.krb Department of System Development, Nuri Solution Co., Yeoido-dong, Yeongdeungpo-ku, Seoul 150-010, Republic of Koreac Department of Electrical Engineering and Computer Science, Korea Advanced Institute of Science and Technology373-1 Guseong-dong, Yuseong-ku, Daejeon 305-701, Republic of KoreaAbstract The field of art psychotherapy is a so called ill-structured paradigm that refers to a vague and unclear problem domain in which there are no clear objective means for finding a solution, and in certain cases the solutions found are inevitably contradictory. The expert system approach provides an appropriate means for finding solutions to such problems encountered in art psychotherapy.Implementation of an expert system to support diagnosis in the field of art psycho- therapy requires extensive knowledge of the characteristics of the drawing, the psycho- logical symptoms of the patient, the relationships between the characteristics of the drawing and the patients symptoms, the diagnosis process, and the methods of know- ledge acquisition. Collaboration among experts from many different fields is needed to provide such knowledge, including art psychotherapists, psychiatrists, psychologists, artists and knowledge engineers. The subjective, ambiguous, inconsistent and sometimes contradictory nature of the huge volume of knowledge involved makes the complex diagnosis process difficult for the field expert. In this chapter, we present an expert system knowledge base to support diagnoses in the field of art psychotherapy and its associated knowledge acquisition method. We provide details of an implementation strategy. In addition, we show how the knowledge base can be represented in terms of ontological framework, using the tool Protégé 2000. We also demonstrate the usefulness and suitability of the proposed approach by applying the system to a real clinical case.Keywords: Art psychotherapy, Knowledge elicitation and acquisition, Ontology, Protégé 2000. 1. Introduction The utility of art psychotherapy using the patients drawings, in which the patient unconsciously divulges his or her inner state of mind or feelings, is widely recognized. This is especially true in the case of children. Children express their innermost emotions through drawings. The value of a child's drawing is worth a thousand words. For example, even children living in poverty or children who are abused by their parents do not criticize their parents openly (Koppitz, 1968). Directly asking family members about the situation is not a useful method (Veltman and Browne, 2001). Traumatized children who have witnessed violence cannot find appropriate words to express their fear, emotions, believing that such expressions are not safe (Malchiodi, 1998). In such cases, a drawing becomes a suitable means of expression, and provides an invaluable step for the treatment of psychological trauma (Gregorian, Azarian, DeMaria, and McDonald, 1996). This is especially true in situations where the childs ability to express himself or herself verbally is underdeveloped (Thomas and Silk, 1990). However, there are numerous difficulties in practice. The relationship between the characteristics of a drawing and the psychological symptoms of a child are extremely complex, since various factors such as the cultural and environmental background of each individual are involved. For example, in the drawings of children who have experienced sexual abuse, many factors, such as their age at the time of sexual abuse, their level of development, the form, period and frequency of abuse, and the relationship with the assailant, are involved to such an extent that it is difficult to find a general relationship between the characteristics of the drawing and the sexual abuse. Therefore, in order to substantiate sexual abuse from a drawing, a sufficient number of characteristics must be simultaneously considered (Malchiodi, 1990). Also, the symptoms that imply sexual abuse are not confined to the characteristics of the drawing alone, but they appear along with other support material on the personal background of the patient (Cohen-Liebman, 1995). Thus, a drawing should not be over-emphasized, and sexual abuse should not be deduced merely from the charac- teristics of the drawing alone (Hibbard, Roghmann, and Hoekelman, 1987). Hence, many studies have questioned the validity of a diagnosis based on the projective function of a drawing, and warn that making a diagnosis based only on a drawing should be done with extreme caution (Lev-Wiesel and Shvero, 2003; Veltman and Browne, 2003). Even if art psychotherapy is performed by a field expert, the diagnosis is based on his or her subjective knowledge, personal experiences and know-how. Thus, the results vary greatly, and in some cases lack consistency or even contradict one another. In an experiment in which abused children were identified through the review of their drawings, Cicchetti and Sparrow (1982) have reported that even art psychotherapists who received sufficient education and training arrived at different conclusions. This is due to the fact that the field of art psycho- therapy is a so called Ill-Structured Paradigm (ISP). An ISP refers to a vague and unclear problem domain in which there are no clear objective means for finding a solution, and in certain cases the solutions found are inevitably contradictory.The expert system approach becomes an appropriate means for finding solutions to such problems encountered in art psychotherapy, which is a typical example of an ISP (Giarratano and Riley, 2005). Furthermore, art psychotherapy requires expert knowledge in numerous fields such as art, psychology, clinical studies and statistics. It is extremely difficult for a single individual to possess such a wide range of expert knowledge. It is nearly impossible for parents or teachers, as non-experts, to perform personally art psychotherapy for their children or pupils. Experts in the field of art psychotherapy are limited in number. Consequently there is a growing need for an expert system approach to overcome such difficulties.Despite the fact that expert systems are in actual use in numerous areas (for example, Kim, Kim, Lee, Kim, and Baik (1992) developed an expert system in law which is another field of ISP), only a few such systems can be found in art psychotherapy. Kim, Ryu, Whang, and Kim (2006) have proposed an expert system approach as a solution to the empirical, heuristic and subjective nature of art psychotherapy methods. Kim, Kim, Lee, Lee, and Yoo (2006) model the complicated mechanism of human diagnosis process as several procedural stages and feedbacks, devise a suitable method of maintaining consistency among numerous decisions derived from the system, and provide the system with a learning facility to improve its intelligence. The methods for automated detection of colors (Kim, Bae, and Lee, in press) and unusual placement types (Kim, Kang, and Kim, 2006) in drawings are proposed. The reason for the lack of expert systems in art psychotherapy may be the reluctance to incor- porate new computer technologies, as would be the case in many non-technical fields.Knowledge related to art psychotherapy is truly variable, broad and abundant. There are numerous books and journals dealing with the different theories as well as the personal experience and know-how of experts. It would be exceedingly difficult to systematically organize this information, because of its subjective, vague, unclear and even contradictory nature. However, the ISP characteristics of art psychotherapy point to the appropriateness of the expert system method. Little research using computer tools and methods seems to have been performed in acquiring, analyzing and systemizing the knowledge used in art psycho- therapy. First of all, the elicitation and acquisition of expert knowledge, experience and know-how are necessary. Then, it is necessary to model the decision process of art psycho- therapy. In this chapter, we attempt to determine the nature of the essential knowledge for art psychotherapy and describe the subsets of this knowledge in a unified object-oriented structure using the ontology concept and Protégé 2000 as a tool. In section 2, an expert system for art psychotherapy is established in the form of a model which consists of four domains comprising the components of art psychotherapy, viz. the characteristics of the drawing, personal environment, psychological symptoms, and mental disorders, and six types of decisions which represent the relationships between these domains, viz. the standard diagnosis, personal diagnosis, psychological diagnosis, and three types of feedback. In addition, we define the drawing tests and the decision process. In section 3, we analyze the knowledge concerning the domains and decisions, and classify and systemize it according to the knowledge structure. In section 4, the appropriate methods of knowledge elicitation and acquisition are presented according to the classification of knowledge. In section 5, we implement the knowledge system into an objected-oriented method, in order to allow for the sharing and reuse of knowledge, using the ontology concept and Protégé 2000 as a tool. Finally, in section 6 we demonstrate the suitability, applicability and possible contribution to art psychotherapy of the proposed system through an actual case study. 2. An Expert System and Its Knowledge Structure An expert system can provide an effective means of overcoming the difficulties encoun- tered in art psychotherapy. For the development of this system, it is crucial to understand the overall process of diagnosis and express it in the form of a model. 2.1. The Necessity In order to examine the possibility of using the expert system, we first identify the difficulties encountered in art psychotherapy. For this purpose, we need to collect the opinions of experts in various fields concerning the difficulties encountered in art psycho- therapy. It is necessary that the experts involved to acknowledge the need for this system and participate actively in the project. The following categories of people are needed for input: Psychological theorists Psychological clinicians Art teachers Art psychotherapists Teachers and parents. The difficulties encountered in art psychotherapy are as follows. The huge volume and variety of knowledge and its inherent inconsistencies and contradictions The complexity, subjectiveness and vagueness of the relationships between the characteristics of the drawing, psychological symptoms and personal environment of the patient The difficulty and subjectiveness associated with discovering the characteristics of a drawing The burden in terms of time and effort The difficulty of use by non-experts. The most difficult problem is that, no matter what the diagnosis may be, the results are not foolproof. Much of the knowledge lacks consistency and even shows contradictions, due to the complexity of the correlations, variations in the research methods, etc. In such cases, the results become clinical, experience-based and subjective rather than theoretical, scientific and objective. An expert system can provide a useful method of solving the problems inherent in an ISP. An expert system can specifically assist the therapist to overcome the difficulties encountered in art psychotherapy, in the following respects. Acquisition and compilation of better knowledge Specification and systematic organization of knowledge Increase in the credibility of the knowledge Interpretations of results of new studies Selection and application of appropriate test (methods) for a drawing Assistance in the process of diagnosis Analysis of the diagnostic process Possible use by non-experts Practical use of the various powerful functions of a computer. Once this system is established, knowledge in numerous areas can be compiled so as to allow for a more appropriate diagnosis and save time and effort. As a result, this system can be valuable for various persons intervening in the process of art psychotherapy. Especially, through the user interface of this system, non-experts such as parents and teachers can easily understand the psychological state of their children or pupils at any time and provide proper guidance according to the information that it provides. The persons involved include: Non-experts such as teachers and parents Psychological theorists Psychological clinicians Art psychotherapists. 2.2. Model of the System In order to construct the expert system, it is first necessary to express the overall process of art psychotherapy in a model. Let us review some examples of the knowledge that we typically handle in art psychotherapy. Knowledge 1 The KSD (Kinetic School Drawing) provides information concerning the adaptation of a child to school and his or her relationships with other children of similar age. Knowledge 2 In a KFD (Kinetic Family Drawing), the index of Effect of Reciprocal Activity in the grading standards of Myer (1978) indicates the degree of satisfaction in family relationships. Knowledge 3 There are cases where a drawing fills up the entire paper, or where a drawing is too large to fit onto a single sheet. The fact that self-expression could not be kept within the limited area of a single piece of paper may imply the possibility of offensiveness, problems concerning control of impulses and behavioral patterns related to such problems. Knowledge 4 Vertical lines whose shape is not clear are known to imply clinically an expression of defiance, arbitrary character, and hostility towards competitors. Knowledge 5 If the score of emotional disorder is higher than a given standard in a psychological test, this implies emotional disorder. Knowledge 6 If the degree of social ability is low in a psychological test, this implies defiance, offensiveness, violence and prejudice. Knowledge 7 If the index of Effect of Reciprocal Activity is high in a KFD despite the fact that the family relationships are not satisfactory, the psychological symptoms that are the result of the evaluation need be reviewed. Knowledge 8 In a KFD, a room which is drawn in a perspective view through the whole screen indicates the severance of communication. However, it implies systemized or organiz- ed consciousness in the case of a Taiwanese child (Wegmann and Vusenbrink, 2000). Knowledge 9 If the scores from the Rorschach Psychological Dissociation Index and the DAP (Draw A Person) Index show a large gap, the psychological dissociation needs to be reviewed (Kent, 1999). Knowledge 10 A child with autism shows introspective, hostile and obsessive psycho- logical symptoms. The above knowledge expressed in natural language can be analyzed to detect the causes (IF) and effects (THEN) and the substantive and declinable words. In this way the substance can easily be extracted. Knowledge 1 presents the usefulness (understanding of personal relationships) of the KSD as a drawing test. Knowledge 2 indicates the relationship between the characteristics of a drawing (the index of Effect of Reciprocal Activity) concerning the KFD test and psychological disorders (social disorders). The KSD and KFD are considered as drawing tests in Knowledge 1 and Knowledge 2. A variety of drawing tests, ranging from DAP, HTP (House-Tree-Person), and FKD (Favorite Kind of Day) to the recently-developed PKFD (Prospective Kinetic Family Drawing) (Taylor, Kymissis and Pressman, 1998), and Circus Drawings (Hanes, 1997), can also be included in other knowledge. Factors such as the utility and characteristics of the drawings are considered according to their tests, and other factors such as their concepts, purposes, conditions, limitations, analysis methods, merits and demerits can be added for consideration. Knowledge 3 indicates the relationship between a characteristic of a drawing (large in size) and the psychological symptoms (offensiveness, problems concerning control of impulses, behavioral patterns). Knowledge 4 also indicates the relationship between a characteristic of a drawing (vertical lines not clear in form) and the psychological symptoms (defiance, arbitrariness, hostility towards others). In Knowledge 3 and Knowledge 4, the factors, the size of a drawing and the form of the lines, are considered as characteristics of a drawing. Factors such as the structure of a drawing, balance, position, color can also be added for consideration. Also, offensiveness, impulsivity, behavior, defiance, arbitrariness, and hostility are considered as psychological symptoms, along with other symptoms such as dejection, fear, happiness, motivation, optimism, violence. Knowledge 5 indicates the relationship between the personal environment (score of emotional disorder) and mental disorders (emotional disorder). Education, family, and friends can be included in the personal environment, and psychological disorders would be such symptoms as mental dissociation, post-traumatic stress, social disorders, etc. Knowledge 6 indicates the relationship between the personal environment (social relations) and psycho- logical symptoms (defiance, offensiveness, violence, prejudice). Knowledge 7 indicates the relationship between the characteristics of a drawing (Index of Effect of Reciprocal Activity) and the personal environment (family relations). Knowledge 8 indicates the relationship between the characteristics of a drawing (a room drawn in perspective) and the psychological symptoms (severance of communication). This relationship is adjusted to represent a different psychological symptom (systemized or organized consciousness) based on the personal envi- ronment (nationality). Knowledge 9 compares mental disorders (Rorschach score) with the characteristics of a drawing (score) and expresses the necessity to review/compare/adjust the diagnosis of the psychological symptoms. Knowledge 10 indicates the relationship between psychological disabilities (autism) and psychological symptoms (seclusion, extremity, introspectiveness, hostility, obsession). By classifying the above knowledge, our art psychotherapy knowledge base consists of four categories of domains with six types of relationships (diagnoses) among these domains. The four domains are (1) characteristics of a drawing, (2) psychological symptoms, (3) personal environment, and (4) mental disorders.The six types of relationships are(1) standard diagnosis which is the relationship between the characteristics of a drawing and the psychological symptoms, (2) personal diagnosis which is the relationship between the personal environment and psychological symptoms, (3) psychological diagnosis which is the relationship between the mental disorders and psychological symptoms, (4) feedback 1 which is the relationship between the personal environment and charac- teristics of a drawing, (5) feedback 2 which is the relationship between the mental disorders and the charac- teristics of a drawing, and (6) feedback 3 which is the relationship between the personal environment and mental disorders.In addition, there is the knowledge pertaining to the selection of the test a child is to draw, and the knowledge concerning the process (or mechanism) of diagnosis. The knowledge of the test selection refers to the stan