THE JOURNAL OF VISUALIZATION AND COMPUTER ANIMATION J. Visual. Comput. Animat. 2003; 14: 269–278 (DOI: 10.1002/vis.324) * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Application of multimedia techniques in the physical rehabilitation of Parkinson’s patients By Antonio Camurri, Barbara Mazzarino,GualtieroVolpe*, Pietro Morasso, Federica Priano and Cristina Re * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
This paper presents and discusses some experiments having the purpose of planning, developing and validating aesthetically resonant environments for different types of sensorimotor impairments which affect Parkinson’s patients. From a technical point of view the aim is to develop a computational open architecture in which it is possible to integrate modules for gesture analysis and recognition and for interactive construction of therapeutic exercises based on multimedia stimulation in real time. The clinical objective is to experiment with a device of sensorimotor stimulation that supports akinesia compensation by controlling movement rhythmic structures in subjects with Parkinson’s disease. The EyesWeb open architecture has been used to analyse patients’ motion and to produce visual feedback during therapy sessions in real time. A pilot study has been conducted on two Parkinson’s disease patients in the framework of the EU IST project CARE-HERE. Copyright # 2003 John Wiley & Sons, Ltd. Received: 1 October 2002; Revised: 1 March 2003 KEY WORDS:
therapy and rehabilitation; Parkinson’s disease; motion analysis; multimedia
techniques
Introduction In the framework of the EU IST project CARE-HERE, we have carried out experiments with the purpose of planning, developing and validating aesthetically resonant environments for different types of sensorimotor impairments which affect Parkinson’s patients. The underlying idea of aesthetic resonance is to give patients a visual and acoustic feedback depending on a qualitative analysis of their (full-body) movement, in order to evoke ludic aspects (and consequently introduce emotional–motivational elements) without the need for the rigid standardization required in typical motion analysis labs, or invasive techniques: the subjects are free to move in 3D with no sensors/markers on body. The scientific community is addressing a growing interest to the investigation and development of therapeutic exercises based on interactive multimedia technologies.1–3 *Correspondence to: Gualtiero Volpe, InfoMus Lab, DIST— Universita` degli Studi di Genova, Viale Causa 13, I-16145, Genova, Italy. E-mail:
[email protected] Contract/grant sponsor: EU IST project CARE HERE.
In the specific context of Parkinson’s disease (PD) Albani and colleagues4 showed that virtual reality (VR) can work as an effective external stimulus in order to explore the motor plans by means of creation of mental images. In this study, a VR environment has been employed reproducing situations related to common daily activities at home, such as eating or using the bathroom. The VR environment has been successfully tested on two women with PD aged 68 and 69 years, and on 10 normal control subjects. Music therapy, including singing, music listening, sharing and discussion of songs, learning to play instruments, song writing, moving to music and participation in music activities, also proved its effectiveness. For example, Pacchetti and colleagues5,6 developed an active music therapy programme combining both motor and emotional rehabilitation for PD patients. Their programme consisted of several group sessions, each based on the correct ordering of listening tasks, exercises and improvisation. Since significant differences have been found among pre- and post-therapy assessments demonstrating a significant improvement in hypokinesia and in daily performances, the authors
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
A. CAMURRI ET AL. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
propose active music therapy as a new method to be included in PD rehabilitation programmes. Our research tackles the rehabilitation task in both its technical and clinical aspects. From a technical point of view the aim is twofold: 1. To develop a computational open architecture in which it is possible (i) to integrate modules for gesture analysis and recognition and for real-time generation of multimedia feedback, (ii) to design dynamic and interactive therapeutic exercises, and (iii) to perform the exercises in real time. 2. To develop algorithms for real-time motion analysis that, despite the limitation given by the lack of onbody markers/sensors, are reliable and precise enough (i) to enable the generation of suitable audio and visual feedbacks in aesthetically resonant environments, and (ii) to allow the therapist to evaluate the progress of the therapy by monitoring the measures associated with a collection of motion features the algorithms provide him. The clinical objective is to experiment with a device of sensorimotor stimulation that supports akinesia compensation by controlling movement rhythmic structures in subjects with PD. The training in recognizing and reproducing these structures can help the subject to control motor tasks that are more complex than those adjusted with simple isochronous signal stimuli. A major fluency in movement recovered with those methods can be transformed in visual or acoustic information produced by the movement itself, which can guide the internal representation of voluntary control. This paper discusses a pilot experiment we carried out in order to test the developed techniques on patients with PD. Patients were chosen on the basis of a previous experimental study developed at the hospital of Arenzano, Italy.7,8 The EyesWeb open software platform (www.eyesweb.org) has been adopted as the basic framework in which the motion analysis techniques have been integrated and the therapeutic exercises developed.
Materials and Methods Subjects and Protocol In a preliminary phase of the study we selected two parkinsonian subjects: a man (aged 67) and a woman (aged 73). The protocol consisted of 12 trials for the male patient during a period of 6 months and six trials for the female patient during a period of 2 months. During
every session, held weekly, the subjects were filmed both during the execution of some simple physical exercises and during audio-visual analyses developed with the EyesWeb open platform.9 In these pilot studies we used both standard material distributed with EyesWeb and new exercises suggested by the physicians of the operative unit, as, for example, an exercise in which the subject is allowed to paint with his own body by moving it in the space. Starting from the considered symptoms (postural instability, gait difficulties and bradykinesia) we selected some simple tasks in order to provide a preliminary taxonomy from which rehabilitation and evaluation exercises can be built. In particular, here we show an example of implementation of the Stand Sit Exercise also used in PD functional evaluation, as in the UPDRS.10
The EyesWeb Open Software Platform The EyesWeb open hardware and software platform9 (www.eyesweb.org) has been adopted for the design, development and real-time performance of physical exercises, the extraction of relevant motion parameters and analysis of the obtained data. EyesWeb is an open hardware and software platform originally conceived for the design and development of real-time music and multimedia applications. It supports the user in experimenting with computational models of non-verbal expressive communication and in mapping gestures from different modalities (e.g. human full-body movement, music) onto multimedia output (e.g. sound, music, visual media). It allows fast development and experiment cycles of interactive performance set-ups by including a visual programming language allowing mapping, at different levels, of movement and audio into integrated music, visual and mobile scenery. EyesWeb is the basic platform of the MEGA (Multisensory Expressive Gesture Applications) IST-20410 EU Project (www.megaproject.org) and has also been adopted in the IST CARE HERE Project on therapy and rehabilitation. EyesWeb is fully available at its website (www.eyesweb.org). Public newsgroups also exist and are managed daily to support the growing EyesWeb community (more than 500 users at the moment), including universities, research institutes and industries. In the particular framework of this study, EyesWeb has been selected since it (i) allows interactive mapping
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
270
J. Visual. Comput. Animat. 2003; 14: 269–278
PHYSICAL REHABILITATION OF PARKINSON’S PATIENTS * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
of motion parameters onto sounds and visual media in a multimedia scenario, (ii) allows integration of novel analysis techniques as new libraries or extensions to existing libraries, (iii) allows fast design, development and testing of novel interactive therapeutic exercises, (iv) can display in real time the analysed physical measurements, and (v) supports different types of sensors (including wireless), one or more cameras, and can be programmed to perform specific analysis of movement in real time. For this last task, the EyesWeb Expressive Gesture Processing and Motion Analysis libraries11,12 have been employed, including software modules for extraction and pre-processing of physical signals (e.g. video from video cameras), and extraction and processing of motion parameters (such as contraction index, directness index, stability index, quantity of motion, pause and motion durations).
Extraction and Processing of Motion Information EyesWeb has been used both to (i) extract and analyse relevant motion features, and (ii) produce aesthetically resonant visual feedback. An archive of recordings from therapy sessions has been collected for preliminary testing and tuning of the developed exercises. Patients have then performed the exercises during therapy sessions with the system working in real time. A layered approach has been followed to move from low-level physical measures (e.g. position, speed, acceleration of body parts) towards descriptors of overall motion features (e.g. motion fluency, directness, impulsiveness) related to PD symptoms. These high-level descriptors are grounded on both the consolidated tradition of biomechanics and on studies by researchers on human movement in the field of performing arts, such as the choreographer and human movement researcher Rudolf Laban and his Theory of Effort,13,14 already used in computational models of human movement.11,12,15,16
Motion Detection and Tracking. Incoming video frames are processed to segment motion and no-motion regions and to detect and obtain information about the motion that is actually occurring. This task is accomplished by means of consolidated computer vision techniques usually employed for real-time analysis and recognition of human motion and activity (see, for example, the temporal templates technique for repre-
sentation and recognition of human movement described by Bobick and Davis.17). It should be noticed that, unlike the research of Bobick and Davis, here we do not aim at detecting or recognizing a specific kind of motion or activity, but rather our goal is to calculate values of cues able to numerically describe qualitative aspects of motion (such as motion fluency). The techniques we use include feature tracking based on the Lucas–Kanade algorithm,18 skin colour tracking to extract positions and trajectories of hands and head, and silhouette motion images (SMIs). An SMI is an image carrying information about variations in the silhouette shape and position in the last few frames. SMIs are inspired to motion–energy images (MEIs) and motion–history images (MHIs).17,19 They differ from MEIs in the fact that the silhouette in the last (more recent) frame is removed from the output image: in this way only motion is considered, while the current posture is skipped. Thus, SMIs can be considered as carrying information about the amount of motion occurring in the last few frames. Information about time is implicit in the image and not explicitly recorded. We also use an extension of SMIs, which takes into account the internal motion in silhouettes, and decomposition of the silhouette into sub-regions (see Figure 1). Information that motion detection and tracking provide to the upper levels is actually encoded in two different forms: positions and trajectories of points on the body (possibly related to specific body parts, e.g. hands, head, feet), and images directly resulting from the processing of the input frames (e.g. human silhouettes, SMIs). Such information is used to extract a collection of motion features. As examples, we describe in the following two such features: quantity of motion (QoM) and contraction index (CI).
Quantity of Motion. The simplest use of an SMI is calculating its area. The result, which we call quantity of motion, can be considered as an overall measure of the amount of detected motion. QoM can be thought as a first, rough approximation of the physical momentum, that is, q ¼ mv, where m is the mass of the moving body and v is its velocity. The shape of the QoM graph is close to the shape of the graphs of velocity of a marker put on a limb. QoM has two problems: the measure depends on the distance from the camera; difficulties emerge when comparing measures from different patients. We solved these problems by scaling the SMI area by the area of the most recent silhouette:
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
271
J. Visual. Comput. Animat. 2003; 14: 269–278
A. CAMURRI ET AL. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Figure 1. (a) Multiple bounding regions. (b) Measure of internal motion using SMIs.
Quantity-of-Motion ¼ AreaðSMI½t; nÞ=AreaðSilhouette½tÞ
In this way, the measure becomes independent from the camera’s distance (in a range depending on the resolution of the video camera), and it is expressed in terms of fractions of the body area that moved. For example, it is possible to say that at instant t a movement corresponding to 2.5% of the total area covered by the silhouette occurred.
Contaction Index. The contraction index is a measure, ranging from 0 to 1, of how the patient’s body is using the space surrounding it. It is inspired by previous studies of the authors on dance performances12 and is related to Laban’s ‘personal space’.14 The algorithm to compute CI combines two different techniques: the individuation of an ellipse approximating the body silhouette and computations based on the bounding region. The former is based on an analogy between image moments and mechanical moments: in this perspective, the three central moments of second order build the components of the inertial tensor of rotation of the silhouette around its centre of gravity: this allows computation of the axes (corresponding to the main inertial axes of the silhouette) of an ellipse that can be considered as an approximation of the silhouette: eccentricity of such an ellipse is related to contraction/ expansion; orientation of the axes is related to the orientation of the body.20 The second technique used
to compute CI relates to the bounding region, i.e. the minimum rectangle surrounding the patient’s body. The algorithm compares the area covered by this rectangle with the area actually covered by the silhouette. Intuitively, if the limbs are fully stretched and not lying along the body, this component of CI will be low, while, if the limbs are kept tightly nearby the body, it will be high (near to 1). While the patient is moving, CI varies continuously. Even if it is used with data from only one camera, its information is still reliable, being almost independent from the distance of the patient from the camera. A use of this feature consists of sampling its values at the end and the beginning of a stretch movement, in order to classify that movement as a contraction or expansion. Suitable multimedia feedbacks can then be mapped onto actions of contraction or expansion depending on the therapeutic goals (e.g. pleasant music can be associated with expansion, if therapy aims at motivating the patient in doing expansion actions).
Motion Segmentation. SMIs have interesting properties: evolution in time of their (normalized) area (what we called quantity of motion) resembles the evolution of velocity of biological motion, which can be roughly described as a sequence of bell-shaped curves (motion bells). In order to segment motion, the sequence of these motion bells and their features, e.g. peak value and duration, have been extracted. A first attempt consists of recognizing phases during which the patient is
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
272
J. Visual. Comput. Animat. 2003; 14: 269–278
PHYSICAL REHABILITATION OF PARKINSON’S PATIENTS * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
rapidly moves from a value near or equal to zero, to a peak and back to zero). On the other hand, a sustained, continuous movement will show a motion bell characterized by a relatively long time duration in which the QoM values have little fluctuations around the average value (i.e. the speed is more or less constant during the movement).
Therapeutic Exercises Figure 2. Motion segmentation. moving (motion phases) and phases during which he or she does not appear to move (pause phases). An empirical threshold has been defined: the patient is considered to be moving if the area of the motion image is greater than 2.5% of the total area of the silhouette. Figure 2 shows motion bells after segmentation: a motion bell characterizes each motion phase.
Fluency and Impulsiveness. Motion segmentation can be considered as a first step towards the analysis of the rhythmic aspects of a patient’s motion. Analysis of the sequence of pause and motion phases and their relative time durations can lead to a first evaluation of how much motion is even and regular, or hesitating. Parameters from pause phases are also extracted to individuate real still positions from active pauses involving low motion (hesitating or oscillating movements). Starting from these data motion fluency and impulsiveness are evaluated. Fluency can be estimated starting from an analysis of the temporal sequence of motion bells. An action (e.g. standing up) performed with frequent stops and restarts (i.e. characterized by a high number of short pause and motion phases) will be less fluent than the same movement performed in a continuous, ‘harmonic’ way (i.e. along a few long motion phases). The hesitating, bounded performance will be characterized by a higher percentage of acceleration and deceleration in the time unit (due to the frequent stops and restarts), a parameter that has been demonstrated to be of relevant importance in motion flow evaluation (see, for example, Zhao,16 where a neural network is used to evaluate Laban’s flow dimension). A first measure of impulsiveness can be obtained from the shape of a motion bell. In fact, since QoM is directly related to the amount of detected movement, a short motion bell having a high peak value will be the result of an impulsive movement (i.e. a movement in which speed
We developed a number of EyesWeb exercises for experiments with PD patients. One of them, for one patient at a time, allows the subject to paint using his or her body. The patient sees himself on a large screen painting in real-time through his motion in the space. Previous work in the performing arts field exists where engagement of the audience is obtained by combined generation of music, sounds and visual media (see, for example, the PAGe—Painting by Aerial Gesture system21). With PAGe the user can interact through an interaction paradigm like the MS Paint software, using his hands while standing in front of a large video screen: the user can select a colour or an action with one hand, then can paint with that colour with the other hand, etc. Our exercise is slightly different: the interaction is based on some of the movement cues described in the previous sections. For example, the colour may depend on fluency; QoM may be associated with intensity of the colour trace; pauses in movement (using the segmentation technique previously described) allow restarting the process and reassigning/adapting interaction mappings. In this way, by a careful choice of colours, e.g. by creating ‘pleasant’ colour associations/mappings with fluent and non-hesitating movements, it is possible to create a sort of visual feedback encouraging improvement of movement in patients. During this exercise the subject looks at the picture painted on the screen and continuously changes it while moving. On another monitor the researcher analyses the parameters and eventually corrects them in order to tune the exercise on the patient’s needs. Figure 3 shows some excerpts from a session with a patient. Several other exercises, involving integrated audio and visual feedback, are currently the subject of investigation. In one of them we analysed fine movements, in particular hand recognition and position (Figure 4). The EyesWeb system allows extracting fine movements in a precise manner and also detects the movement intention by recognizing arm movement in order to be easily suited to severely impaired persons.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
273
J. Visual. Comput. Animat. 2003; 14: 269–278
A. CAMURRI ET AL. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Figure 3. The painting exercise. We also analysed simple tasks in order to provide a preliminary taxonomy from which we may design rehabilitation and evaluation exercises. In particular, here we show an example of implementation of the stand–sit exercise also used in PD functional evaluation as in the UPDRS (Figure 5). The challenge and the aim of our preliminary study was to evaluate if such kinds of systems are able to generate aesthetically resonant feedback in PD patients,
providing the emotional/motivational involvement to allow them at least partially to overcome their disease. This will be the subject of a future clinical study. At the moment, however, an analysis of a questionnaire about Quality of Life that the patients filled in after each therapy session (12 trials for the male patient during a period of 6 months and six trials for the female patient during a period of 2 months) demonstrated an overall positive trend and an average increment of
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
274
J. Visual. Comput. Animat. 2003; 14: 269–278
PHYSICAL REHABILITATION OF PARKINSON’S PATIENTS * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Figure 4. Analysis of hand movements.
Figure 5. Stand–sit exercise. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
275
J. Visual. Comput. Animat. 2003; 14: 269–278
A. CAMURRI ET AL. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
4. Albani G, Pignatti R, Bertella L, Priano L, Semenza C, Molinari E, Riva G, Mauro A. Common daily activities in the virtual environment: a preliminary study in parkinsonian patients. Neurological Sciences 2002; 23(Suppl. 2): S49–S50. 5. Pacchetti C, Aglieri R, Mancini F, Martignoni E, Nappi G. Active music therapy and Parkinson’s disease: methods. Functional Neurology 1998; 13: 57–67. 6. Pacchetti C, Aglieri R, Mancini F, Martignoni E, Nappi G. Active music therapy in Parkinson’s disease: an integrative method for motor and emotional rehabilitation. Psychosomatic Medicine 2000; 62: 386–393. 7. Morasso P, Baratto L, et al. A new method for the evaluation of postural stability in Parkinson’s disease. In Proceedings of Medical and Biological Engineering and Computation 1999; 37(Suppl. 2): 822–823. 8. Re C, Baratto L, et al. Analysis of movement control strategies in Parkinson’s disease (Abstract). Gait and Posture 2001; 13: 158. 9. Camurri A, Coletta P, Peri M, Ricchetti M, Ricci A, Trocca R, Volpe G. A real-time platform for interactive dance and music systems. In Proceedings of the International Computer Music Conference ICMC2000, Berlin, 2000. 10. Fahn S, Elton R. Unified Parkinson’s disease rating scale. In Recent Developments in Parkinson’s Disease, vol 2, Fahn S, Marsden CD, Calne DB, Goldstein M (eds). Macmillan Health Care Information, 1987. pp. 153–163, 213–304. 11. Camurri A, Lagerlof G, Volpe G. Recognizing emotion from dance movement: comparison of spectator recognition and automated techniques. International Journal of Human–Computer Studies; at press. 12. Camurri A, Trocca R, Volpe G. Real-time analysis of expressive cues in human movement. In Proceedings of the International Computer Music Conference ICMC2002, Gothenburg, 2002. 13. Laban R, Lawrence FC. Effort. Macdonald & Evans: London, 1947. 14. Laban R. Modern Educational Dance. Macdonald & Evans: London, 1963. 15. Camurri A, Hashimoto S, Ricchetti M, Trocca R, Suzuki K, Volpe G. EyesWeb: toward gesture and affect recognition in dance/music interactive systems. Computer Music Journal 2000; 24: 57–69. 16. Zhao L. Synthesis and acquisition of Laban movement analysis: qualitative parameters for communicative gestures. PhD dissertation, University of Pennsylvania, 2001. 17. Bobick AF, Davis J. The recognition of human movement using temporal templates. IEEE Transactions on Pattern Analysis and Machine Intelligence 2001; 23(3): 257–267. 18. Lucas B, Kanade T. An iterative image registration technique with an application to stereo vision. In Proceedings of the International Joint Conference on Artificial Intelligence, 1981. 19. Bradsky G, Davis J. Motion segmentation and pose recognition with motion history gradients. Machine Vision and Applications 2002; 13: 174–184. 20. Kilian J. Simple Image Analysis By Moments. OpenCV library documentation, 2001. 21. Tarabella L, Bertini G. Wireless technology in gesture controlled computer generated music. In Proceedings of the Workshop on Current Research Directions in Computer Music, Barcelona, 2001.
patients’ satisfaction that moved from 33% at the beginning of the therapy to 60% after it.
Discussion The pilot experiments demonstrate the feasibility of the project: patients have understood in a natural way the dynamics of the proposed resonant environments and they obtained benefits from them, if nothing else, from a motivational point of view. The acquisitions made during the interactive exercises highlighted a general improvement of movement fluency, and appropriate EyesWeb patches for automatic analysis will be developed. Of course, this is only a preliminary study because it could be applied only to a very limited population of patients. However, the experience was essential for understanding the practical problems that could allow the organization of a clinical trial in the near future. In particular, we defined the environmental requirements, the acquiring protocol and some guidelines for designing the proper exercises for analysis and improvement of movement rhythmic structures. Overall, the EyesWeb architecture demonstrated great versatility, giving a new possibility to the physician to focus exercises and analyses on the desired rehabilitative target and to evaluate the efficacy of the therapy. ACKNOWLEDGEMENTS
The work described in this paper has been partially supported by the EU IST project CARE HERE. We thank our colleagues Riccardo Trocca and Matteo Ricchetti at the InfoMus Lab, physicians Luigi Baratto and Marcello Farinelli, Psiche Giannoni and Emanuela Cervetto, Grazia Ornato, Luciana Tabbone and Alessandro Tanzini at CBC for their contribution to this work. We also thank the EyesWeb development team (Paolo Coletta, Massimiliano Peri and Andrea Ricci) for their support.
References 1. Bianchi Bandinelli R, Saba A. Proposals and solutions for autonomy of the disabled and elderly. Studies in Health Technology and Informatics 1998; 48: 140–144. 2. Wong SK, Tam SF. Effectiveness of a multimedia programme and therapist-instructed training for children with autism. International Journal of Rehabilitation Research 2001; 24(4): 269–278. 3. Brown DJ, Powell HM, Battersby S, Lewis J, Shopland N, Yazdanparast M. Design guidelines for interactive multimedia learning environments to promote social inclusion. Disability and Rehabilitation 2002; 24(11–12): 587–597.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
276
J. Visual. Comput. Animat. 2003; 14: 269–278
PHYSICAL REHABILITATION OF PARKINSON’S PATIENTS * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
sive gesture in human full-body movement. Now she is a PhD Student in Computer Science Engineering and she works at the DIST—InfoMus Lab of the University of Genova, directed by Prof. Antonio Camurri. She works on the modeling and real-time analysis and synthesis of expressive content in dance and human movement.
Authors’ biographies: Antonio Camurri is Associate Professor at University of Genova where he teaches the courses of ‘Software engineering’ and ‘Multimedia Systems’ (Computer Engineering curriculum). He is the founder and scientific director of the Laboratorio di Informatica Musicale at DIST, member of the Board of Directors of AIMI (Italian Association for Musical Informatics), member of the Executive Committee (ExCom) of the IEEE CS Technical Committee on Computer Generated Music, founding member of the Italian Association for Artificial Intelligence (AI*IA), Associate Editor of the international Journal of New Music Research (Swets and Zeitlinger). He is author of more than 70 scientific international publications, and served in the scientific committees of several international conferences. He is owner of patents on software and computer systems, algorithms and multimedia systems. He is Chairman of the following scientific events: ‘First Intl Workshop on Kansei—The Technology of Emotion’, Genova, October 1997, AIMI and IEEE CS. Track on ‘Kansei Information Processing’ at the IEEE Intl Conf SMC’98, San Diego (3 Scientific Sessions). Track on ‘Kansei Information Processing’ at the IEEE Intl Conf SMC’99, Tokyo, co-chair of the 5th Intl. Gesture Workshop, and Guest Editor of a special issue of IEEE Multimedia J. on ‘Multi-sensory communication and experience through multimedia’ (Jul–Sep 2004).
Barbara Mazzarino was born in Genova, Italy, on the 30th of July 1976. She received her Laurea degree in Computer Engineering from the University of Genova on April 23rd 2002, with a thesis on analysis of expres-
Gualtiero Volpe computer engineer, was born in Genova on March 24th, 1974. He received his degree in computer engineering from the University of Genova in 1999. He will discuss his PhD dissertation in June 2003. His research interests include intelligent and expressive human-machine interaction, modeling and real-time analysis and synthesis of expressive content in music and dance, KANSEI information processing and expressive gesture communication. At the moment he is researcher at the DIST—InfoMus Lab at the University of Genova. He is co-chair of the 5th Intl. Gesture Workshop (Genova, April 15–17, 2003).
Pietro G. Morasso was born in Genova, Italy, on April 30, 1944. In 1968 he received his Laurea degree in Electronic Engineering cum laude from the University of Genova. He has been associated with the MIT Department of Psychology, Cambridge, Mass. USA, in the Neurophsysiological laboratory of Prof. Emilio Bizzi as a post-doctoral fellow (1970–1972), Fullbright fellow 1973, vsiting professor 1978, 1979, 1980. He had different permanent positions in the Engineering Faculty of the University of Genova since 1970. Currently he is full professore of Anthropomorphic Robotics and chairman of the Laurea Programme in Biomedical Engineering.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
277
J. Visual. Comput. Animat. 2003; 14: 269–278
A. CAMURRI ET AL. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
He is founder and scientific director of the Bioiengineering Center at the Rehabilitation Hospital La Colletta in Arenzano since 1995. His scientific interests include the following topics: analysis of the motor control system; computational neuroscience; neuroengineering; anthropomorphic robotics; rehabilitation engineering. He is author/coauthor of 6 books, over 300 papers, and two patents (one on robotic navigation and one on lasertherapy).
Federica Priano was born in Genoa on October 25, 1966. She received her degree in Psychology at the University of Padua in 1991 and was licensed as a Psychotherapist in 1999. In 1997, she received her PhD in ‘Psychodynamic and Neurophysiology’ at the University of Genoa discussing a thesis on ‘Cognitive Rehabilitation’. She found the Laboratory of Clinical and Rehabilitation Neuropsychology in the Rehabilitation Department (Chairman Dr. Baratto) of ‘La Colletta’ Hospital in Arenzano (Genoa) where she dealt with Neuropsychological Assessments and Cognitive Rehabilitation programs to patients with neurological diseases (stroke, Parkinson’s disease, Multiple Sclerosis, traumatic brain injury, tumors, dementia, etc). She also cooperates with the Bioengineering Center at the same Hospital. She is author/co-author of 19 scientific publications.
Cristina Re was born in Albenga, Italy, in 1971. She received her degree in electronic engineering from University of Genoa in 1999 with the thesis: ‘Analysis and classification of posturographic parameters by neural networks computation’ supervised by prof P. G. Morasso. Currently she is PhD student in Bioengineering and Bioelectronic at University of Genoa and attends her researches in the staff of prof. P. G. Morasso and MD L. Baratto at the Centre of Bioengineering—‘La Colletta’ hospital—Arenzano. Her research interests include Posturographic and Surface Electromiographic Data Analysis, Motion Analysis, Neural Networks. She has also been involved in the project CAREHERE financed by the European Community.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Copyright # 2003 John Wiley & Sons, Ltd.
278
J. Visual. Comput. Animat. 2003; 14: 269–278
