Inter-subjectivity is fundamentally different in persons with autism compared to non-autistic persons. The ability to perceive others, to feel with them, or to take their perspective, the ability to socialize, to make conversation, small talk, to relate to others, or to establish bonding is impacted in autism. For non-autistic persons, social interaction with persons with autism is often experienced as awkward or strange [1-4].

Dance movement therapy (DMT) addresses the nonverbal abilities to relate non-verbally, and to feel comfortable when relating nonverbally. It further addresses the ability to sense one’s own body, to distinguish self from other, and to feel comfortable in one’s own body and in relationships with others. From an embodiment perspective, DMT strengthens the resources of persons with autism, improves their body perception and body image, and increases their well-being, their social competences and their empathy. A pilot study of DMT with young adults with ASD [5] based on a series of mirroring interventions showed that body awareness, social competence, well-being/affect, and differentiation of self and other increased significantly after dance movement therapy, while empathy did not increase significantly.

Based on this pilot study, the TESIS-group is presently conducting a multi-center RCT-trial on dance movement therapy with an ASD population. While the quantitative investigation on empathy and other variables of interest is still on-going, we investigated body image changes using a qualitative projective approach, the body image sculpture test (KST; [1]; see method section) (Figure 1).

Figure 1: Interviewer: “What of the sculpture reminds you of yourself?” Participant: “Feelings of inner deformation, of pathological self-less-ness.”

Body image

Body image is a concept pertaining to the conscious representation of one’s body in the mind [6-10]. Generally, we distinguish between body image and body schema. Body schema describes the widely implicit (or unconscious) representations of the extension and functions of the body (from bottom-up sensory input), and body image encompasses the pre-conscious and conscious representations and evaluations of the body (also including top-down processes). What can be said about their relation?

In body psychotherapy, there is much agreement on “the self being primarily and foremost a bodily self” [11]. However, there is much conceptual confusion around the terms of body image and body schema. Schilder [9,12], who introduced the terms into clinical practice, used both terms in the same sense. Joraschky et al. [8], sorting out the different uses, approximates a definition of body image as an aspect of our subjective body experience. As a cognitive concept that constitutes, and differentiates itself individually, it develops based on experience and in close parallel to speech acquisition. Joraschky et al. [8] conception of body image is made up of three change-sensitive components: body knowledge, body fantasies, and body concepts. Functionally, it symbolizes the individual reference to the body in social contexts and forms the basis of important intra-individual functions such as body boundaries, self/identity, and gender role identity. It contains the body schema. The body schema according to Joraschky et al. [8] develops perceptively and is ontologically earlier than the body image. It refers to representations of physical extensions and functions of the body. (Joraschky et al. [8] translated by the authors). However, Joraschky et al. [8]) also emphasizes that no descriptive terminology is presently adequate to cover the complex Gestalt of the entire subjective body experience.

Shaun Gallagher makes a point of distinguishing between body image and body schema. Body schema describes the “non-conscious performance of the body” [5] and refers to its physiological and behavioral functioning, its posture and habits as it interacts with, and is simultaneously part of, its environment. The body image, then, is the perceptual, cognitive, and emotional conception of the body as it is appears in consciousness. In awareness, the body as an image lives not as an integrated whole (as the body the schema does), but as a reflection that is abstract and disintegrated. This reflection is dependent on a multitude of factors, of which the body schema forms only a part [5].

The intra-individual functions of the body image and its relation to inter-subjectivity are covered in recent mirror neuron research [13]. In a seminal article “The case for motor involvement in perceiving conspecifics”, Wilson and Knoblich [14] point out that covert imitation functions as part of a perceptual emulator, using implicit knowledge of one’s own body mechanics as a mental model to track another person’s actions in real time.

In ASD with its many sensory integration aspects, it is likely that the embodied self is profoundly affected (cf. [15,16]). Given the potential implication of body image for inter-subjectivity [13,14], as a core problem in autism, body image in autism seems a topic worthwhile investigating.

Our study

We hypothesized that participation in ten dance movement therapy group sessions would improve the body image (as defined by Gallagher [6] and Joraschky et al. [8], thus including the body schema) of young participants with ASD measured with the Body-Image-Sculpture-Test (KST; von Arnim [1,17]). The KST measures individual body image, yet with an inclusive understanding of aspects of inter-corporeity [18], on the basis of the constant reciprocal influence of our own and others’ body image [2].

The intervention

“Critically, what makes the “mirroring” idea and related phenomena appealing is the possibility that they reveal a non-representational relation to others”

(Carr and Winkielman [13], referring to Gallagher [19], Hutto [20] and Sinigaglia [21]).

On the basis of the pioneering work of Adler [22], we employed mirroring in movement that we manualized for the purpose of a bigger randomized study. Every session consisted of the same sequence of mirroring exercises and a verbal processing part.

(a) Warm-Up (about 10 min): We used the Chace-Circle ([23]), a circle formation where the therapist picks up elements of each participant and asks the group to try them out (“can we all do what Mr. X does?”, “can we all be with Mrs. Y?”) and playfully change them (“can we make this bigger/smaller/louder/softer,” etc.). The Chase-circle creates an atmosphere of being seen and accepted as one is, and a secure therapeutic space, where participants can experiment and express their thoughts and feelings [23]. After the warm-up and under the condition that an atmosphere of acceptance had been created, all participants split into dyads.

(b) Dyadic movement part (about 20 min): A dyad consisted of either two participants or one therapist/assistant and one participant. Each participant had the opportunity to choose his or her preferred partner. First, one participant was asked to lead, the other followed; then upon the second song the partner was asked to lead and the first participant followed; and upon the third song both were asked to move freely but to always stay in contact with each other, no matter whether they were moving close to each other or at the opposite sides of the room. We emphasized that it was not important during mirroring that each person exactly mirrors the shapes of the other person’s movement, but that it was important that their movements reflect the quality of the other’s movement, genuinely trying to ‘be with them’. For the dyadic mirroring part, we used a mix of slower and faster, short pieces of music (each at maximum 3 min). This free dancing part also ensured that participants had the opportunity to freely choose the level of mirroring them preferred [24].

(c) Baum-circle (about 20 min): Next, all participants came back together in a circle. The movement part was continued with a “Baum-circle” [25]). For this part of the session, participants were encouraged to bring a piece of their own music, which was meaningful to them. Volunteers initiated movement to his/her selected music located as a regular part of the circle, being asked to focus on the expression of his or her feelings, while not paying too much attention to the others. All other participants were asked to follow in the same kinesthetically attuned way they did before in the dyads. The Baum-circle aims to establish rapport and empathy in the participants using kinesthetic attunement and emotional contagion [25,26]. Up to four volunteers initiated movement in the Baum-circle in each session. The aims were to ‘feel with’/empathize, understand (“to go in the other’s shoes”), and respect the other. Participants’ use of this technique - originally developed for trauma patients – demonstrated that it was suitable for persons with ASD.

(d) Verbal processing part (about 10 min): Finally, all participants sat down to reflect on the session, moderated by the therapist. In this context, the participants could verbally express their feelings and their opinion regarding the session. The therapist first encouraged the participants who initiated an improvisation in the Baum-circle to verbalize how it was to move, what they wanted to express, and how it felt to be reflected by the other group members. Then, the other participants were asked about their perceptions and feelings when they moved with the person. The aim was to provide and receive feedback suited to increase body-awareness, self-awareness, self-other awareness, empathy, and social skills, and to verbalize the non-verbal experiences and feelings.

The entire DMT session lasted 60 min (for a more detailed description see Koch et al. [5]).

Sample

The sample consisted of ten (N=10) young adults with ASD (8 men, 2 women; mean age=23.1, SD=8.7; range 17-43) at a professional rehabilitation institution in the Heidelberg region. Most participants belonged to either high-functioning or Asperger Symptom groups. Their main consultant rated their degree of impairment (autistic symptoms level) as “light” in one participant, “medium” in five participants and “severe” in four participants. They were all part of the rehabilitation program for professional training at SALO and Partner GmbH in Southwest Germany, where they were part of the “AuReA”-program (Autismus – Rehabilitation – Arbeit), a specific training program for persons with Autism with a focus on integration into the labour market.

Participants were required to be age 16 or older, and able to perform movement for 1 h in a standing position. They did not receive any other form of psychotherapy while participating in the DMT study. The manualized intervention consisted of the mirroring-based DMT techniques described above and is fully described in Koch et al. [5].

Procedure

The KST was administered by one of the therapy helpers before the second therapy session; it was administered again after the last therapy session by the same therapy helper. Participants’ task was to “close their eyes and form a human figure”; the complete instructions are in Box 1. Participants took between 10 and 30 min each, indicating that the times for completion varied greatly. Participants sat at a single table, with the pre-prepared clay, a box, a towel, and the questionnaire. Participants were given the choice to either model the human figure with their eyes closed or to use the box covered with the towel in order to form the figure without visual feedback. Visual concealment is important to ensure that subjects go from the kinesthetically felt sense of the body rather than using visual comparison. When they were done modeling, participants completed the questionnaire on their thoughts, feelings and bodily sensations in relation to their sculpture. The questionnaire was used with the goal to provide participants the opportunity to reflect about the process and result of the sculpting without having to speak in front of the group (which often only yields stereotypical answers). The KST was conducted in small groups of participants (four at the most) to provide a safe space and the opportunity for personal communicating.

Instruments: The KST

The Body Image Sculpture Test (Körperbild-Skulptur-Test, KST; [1,2]; see also [27-29]) is a projective clay work task in which the participants are blindfolded and then asked to model a human figure (sculpture) from a piece of clay (of approx. 500 g weight). For this study, participants were merely asked ‘to keep their eyes closed’ in order to leave them the maximum control possible.

Within the body work tradition in Germany, the earliest note we have of the use of a body sculpture test for embodying aspects of the body image is from the work of Alexander [27], the founder of Eutonie, a body work technique. She used it for a spatial image of the body. The tradition was picked up amongst others by Schubert [28] and Wadepuhl and Wadepuhl [29] who each refined the technique. For our study we used von Arnim’s version [1,17]. In this tradition, the technique is informed by psychodynamic theory and mostly applied in clinical contexts.

The standardized instructions were the following:

We used the coding manual of von Arnim [17] to categorize the figures. It contains ratings of five dimensions on a 3-point Likert-scale (0-2; poor to elaborate): connectedness, proportion, dimensionality, surface quality and completeness. All different body parts were coded and then the median of each dimension was computed. Single coding dimensions:

Connectedness: Was the figure formed from one piece, how well is the figure connected, were the transitions between body parts roughly or thoroughly joined? The more connected the higher the score.

Proportion: Taking the properties of the torso as orientation, the rest of the figure is set in relation: are the proportions realistic? To account for left-right symmetry, pairs of body parts are coded separately. Proportion accounts for two-dimensional relations. The more realistic the proportions, the higher the score.

Dimensionality: Is the back of the sculpture flat? Does the figure have a realistic volume? Each body part is scored for threedimensionality. The more realistic the three-dimensionality, the higher the score.

Surface quality: How is the surface worked (rough vs. thorough)? Are there deformations? The less roughness and deformation, the higher the score.

Completeness: How complete is the figure? Seven body parts are rated regarding completeness, paired body parts receive single scores. The more complete the figure, the higher the score.

Interrater-reliability of dimensional ratings

The therapy helper, a female student, rated the entire material and a second female observer rated 50% of the material to determine the observer agreement. Sculptures for the second rater were selected in a random draw. The intra-class-correlation [30] was computed with a resulting intra-class-coefficient of rICC=0.70, indicating a good agreement [31].

Questionnaire

The questionnaire consisted of fifteen questions (see Appendix B) and served the purpose of including participants’ reflections on the process and outcome of forming their sculpture.

Analysis

The analysis was done with simple pre-posttest comparisons (within-subject) by t-tests. Results of the interviews were reported descriptively. The alpha-level was set at 0.05.

Main results

From pre- to post-test the dimension of „proportions “(prop) showed a significant increase t(9)=-4.47, p=0.002, as did “completion” (com) t(9)=-4.45, p=0.002, „connectedness “(con) t(9)=-3.48, p=0.007, “dimensionality” (dim) t(9)=-2.94, p=0.017, and “surface quality” (surf) t(9)=-2.53, p=0.032. Means are reported in Table 1.

Table 1: Means of KST-Ratings; Note. In a sample of 10 participants with ASD receiving ten weeks of dance movement therapy (mirroring manual), 9 formed a human figure and showed improvement in body image – as operationalized by the connectedness, dimensionality, proportions, completeness and surface quality of the figure -- at post-test. The inter-rater-reliability was good with rICC=0.70 (Wirtz and Caspar [31]).

Interviews

While participants’ felt more insecure the first time around, interviews yielded many positive statements about the sculptures, but also statements reporting a „loss of control while forming“ or an „inner feeling of insecurity or inner suffocation“. All participants were able to name positive aspects of their figure. One participant stated at posttest to “feel more complete”. Many participants said they had imagined their sculpture just the way it was; two had imagined it to be smaller. Only two participants answered that they would not correct any aspect of their sculpture. In response to the question “what of your sculpture reminds you of yourself?”, some said the “form of a human body”. One participant said: “Feelings of inner deformation, of pathological self-lessness” and was then talked to in private to assess and address the severity behind this utterance. Participants took 10-30 min to model their sculptures (we did not standardize to 10 min, but left more freedom for participants to work in their own time).

Sculptures

All sculptures showed improvements at post-test. The figure at pretest is presented on the left and the figure at post-test on the right side to enable direct comparison (Figures 2-11).

Figure 2: Subject 1 - Post-test (right side) shows an increase of proportion, completion, dimensionality, connectedness, and surface quality.

Figure 3: Subject 2 - Post-test (right side) shows an increase of dimensionality and elaboration (completion). Note the change in body posture from lying to sitting.

Figure 4: Subject 3 - Post-test (right side) shows an increase of proportion, completion, dimensionality, and connectedness; but also three arms (distortion).

Figure 5: Subject 4 - Post-test (right side) shows an increase of proportion, dimensionality, connectedness, and surface quality.

Figure 6: Subject 5 - Post-test (right side) shows an increase of proportion, completion, dimensionality, connectedness, and surface quality. Note the higher degree of elaboration/details of the face, body (hands and feet), and clothing at t2.

Figure 7 (excluded): Subject 6 - Post-test (right side) shows an increase of proportion, completion, dimensionality, and surface quality; the participant was not able to form a human figure, which was explicitly instructed, but formed a mascot of his girlfriend - a little sheep - at both times. Since he was the only one to form a non-human sculpture despite the instructions, and also because we could not say for sure what the sequence of the two sculptures was, we excluded subject 6 from the final analysis. Nevertheless, it remains interesting that he formed a sheep.

Figure 8: Subject 7 - Post-test (right side) shows an increase of proportion, completion, and dimensionality.

Figure 9: Subject 8 - Post-test (right side) shows an increase in proportion, completion, dimensionality, and connectedness.

Figure 10: Subject 8 – Post-test (right side) shows a slight increase in connectedness and surface quality.

Figure 11: Subject 10 – Post-test (right side) shows mainly an increase in connectedness.

In many figures hands, feet, and other elaborated details were only found at post-test.

Results suggest body image improvement in young adults with ASD on all dimensions of the KST after a dance movement therapy intervention based on therapeutic mirroring: connectedness, proportion, dimensionality, surface quality, and completeness. In sum, the sculptures have a low to medium structural level (expert judgment). However, there are no normative values for the KST yet.

A replication of the study with a control group and a bigger sample in a randomized design is recommended. Validation of the KST in controlled studies with more N is needed. Given the effect sizes of the present study, N=64 would be needed to fulfil the requirements of power analysis.

Is it a coincidence that all sculptures improved? Since we had no control group, we cannot be sure and recommend repeating the study with an active control group. This is particularly necessary in order to exclude the possibility that sculptures improved from t1 to t2, merely because participants were more confident with the now more familiar task. Familiarity of situations in an important safety feature for all persons with ASD.

Can the change in the sculptures be causally related to the dance movement therapy intervention of therapeutic mirroring? The answer to this question again requires a control group to draw a qualified conclusion about causal and other factors. An active control group might be an exercise group or another arts therapies group. Future studies should also use a second measure of body image, such as a suitable self-report measure. This is important, because the KST is a projective test and thus always subject to pre-assumptions and interpretations that may not be shared in the entire scientific community. There remains an interpretative leap in the questions of how the body image of the participants is expressed in the sculptures.

Future studies could further use the interview more extensively in order to allow participants to convey more information about their subjective body experience in a narrative form. This way, we could obtain more differentiated, qualitative information about the body image in autism. This could, for example, be accomplished by integrating parts of the EASE interviews of Parnas et al. [32], originally developed for schizophrenic patients, with the KST interview and applying it to participants with ASD.

“Real life skill is better understood, if the sensorimotor origin of cognition is not abandoned” Thelen [33].

In addition to the treatment of autism with cognitive-behavioral methods [34], there is a recent trend to include interventions on the sensorimotor level [5,35]. Sensorimotor interventions reveal an important access to persons with autism as a resource-oriented approach, affecting particularly the body-self and the improvement of social skills [5].

Nonverbal approaches such as dance or music therapy are suited to address the needs and difficulties in autism [36]. They could reach deeper than verbal approaches, promoting nonverbal communication and expression as a pathway to the improvement of smoother dialoguing and more confident conversing. Reaching into preverbal mirroring modalities helps persons with autism to improve their body image, body awareness, awareness of self-other distinction, selfexperienced social competence and well-being (particularly relaxation; Koch et al. [5]. Via the nonverbal, the body can activate resources that cannot be accessed by verbal or cognitive means. The body is a reservoir of resources - if we start diving into it, it is surprising what can happen (cf. Jones [37]). While further research is needed to support these preliminary findings, these first results support the assumption that dance therapy can stimulate the development of body image, furthering body awareness, identity formation, and inter-subjectivity and thus addressing the core symptoms of autism.