You are in: eMedicine Specialties > Neurology > Computer Applications in Neurology Virtual Reality Biofeedback in Chronic Pain and PsychiatryArticle Last Updated: Dec 8, 2006AUTHOR AND EDITOR INFORMATIONSection 1 of 10
  Author: Morris Steffin, MD, Chief Science Officer, Virtual Reality Neurotech Lab Editors: Marion Priscilla Short, MD, Assistant Professor, Departments of Neurology, Pediatrics, and Pathology, University of Chicago Hospitals and Clinics; Francisco Talavera, PharmD, PhD, Senior Pharmacy Editor, eMedicine; James H Halsey, MD, Professor, Department of Neurology, University of Alabama Medical Center; Matthew J Baker, MD, Consulting Staff, Collier Neurologic Specialists, Naples Community Hospital; Nicholas Y Lorenzo, MD, Chief Editor, eMedicine Neurology; Consulting Staff, Neurology Specialists and Consultants Author and Editor Disclosure Synonyms and related keywords: VR, psychiatric disorders, pain disorders, pain physiology, biofeedback in pain treatment, virtual reality applications in biofeedback, psychiatric applications of biofeedback INTRODUCTIONSection 2 of 10
  Active modification of nociceptive perceptual states through biofeedback has been of considerable benefit in reducing patient discomfort related to several disease processes. Virtual reality (VR) allows substantial amplification of traditional biofeedback efficacy by completely immersing the patient in the process. To accomplish this, an immersive, active patient environment is generated to achieve more effective pain and anxiety reduction in the context of chronic pain, headache, and anxiety states. While the application of VR in this context is new, the neuropsychological basis for its effects has been well studied in a variety of pain and psychiatric disorders. PAIN PHYSIOLOGY: WHY SHOULD VIRTUAL REALITY WORK?Section 3 of 10
Recent investigations with functional MRI studies have reinforced the notion that pain perception and affective response are dependent on several topographically diverse regions. In addition to primary and supplementary sensory areas, the insula also is involved, although overlap of nociceptive and nonnociceptive stimuli can produce ambiguity in the responses. Cingulate cortex has been implicated in pain processing, and also in general attention mechanisms involved with demanding tasks. In determining pain response levels, central processing appears more important than simple rate of firing of spinal neurons over regions of activation at segmental levels. A more recent positron emission tomography (PET) approach has defined a clearer role for several regions. Gating function determining pain threshold appears to be located in anterior cingulate cortex and inferior frontal areas along with the thalamus. Pain intensity determination appears to involve periventricular gray and posterior cingulate cortex activation. The actual affective (unpleasantness) component appears to be related to activation of the posterior sections of the anterior cingulate cortex. At all levels in the pain pathway, processing of discriminative and affective pain components appears to be separate, and modulation of processing encompasses bidirectional (ascending and descending) control systems. Descending pathways arising in cortex, thalamus, and brain stem affect processing dynamics at the spinal levels. Particularly strong bulbar descending inhibitory effects on dorsal horn nociceptive neurons have been described. Basal ganglia responses also may modulate pain perception. Modulation of pain perception by stimulation and attention In view of these physiological mechanisms, that redirection of attention can alter pain perception markedly is not surprising. For example, such redirection has been shown to modify nociceptive responses in neuronal populations in the dorsal medulla at the same time that pain unpleasantness characteristics are modified. Such attentional effects, arising from hypnosis anesthesia, can suppress spinal reflexes. In contradistinction to inhibitory mechanisms, central mechanisms are also capable of producing pain hyperexcitability through neuroplastic mechanisms preconditioned by frequent noxious stimuli. Furthermore, hypnosis analgesia and acupuncture analgesia appear to affect pain perception without participation of the endorphin system. This suggests that some of these modulatory systems may be recruited during such procedures. Acupuncture has been shown to provide pain relief for diabetic neuropathy. At the same time, transcutaneous electrical nerve stimulation (TENS) has been shown in animal studies to exert a direct, segmental inhibitory influence on nociceptive dorsal horn neurons. The multiplicity of activation at spinal, brainstem, diencephalic, and cerebral levels, coupled with the multidirectional, modulatory influences on pain processing signals, provide, at least conceptually, ample substrate for the effects of biofeedback described below. Two major effects on pain perception are reasonable, but difficult to quantify, on the basis of these described mechanisms. A pronounced time integration effect on pain perception is based on the level of pain and its duration. Thus, a moderate, nagging, persistent pain may wear the patient down as much as a brief but more severe pain. Periods of freedom from pain allow this effect to decay, thus resetting pain perception levels. The attention-diverting amelioration of pain is bidirectional (ie, focusing attention on the pain makes it worse). This is especially evident when the patient's anxiety concerning the cause of pain is increased. For example, a patient with persistent low-level abdominal pain will perceive it as much less severe if reassured that the cause is benign and as more severe if informed that the cause might be a malignant condition. A more dramatic example occurs in the course of sports or other physical injury environments. Often serious injuries are not perceived as painful until after the condition of high motivation (eg, scoring the touchdown, fleeing an assailant) has passed. STANDARD BIOFEEDBACK APPLICATIONS IN PAIN TREATMENTSection 4 of 10
Multiple mechanisms for pain modulation are based on stimulus input. The rationale behind biofeedback for pain treatment is that diverting or competing stimuli may improve both immediate pain levels during the biofeedback procedure and during intervals between exposures. However, applications of biofeedback techniques have been largely empirical. With VR, the capacity to increase levels of immersion is likely to increase the power of the technique. Before considering such approaches, certain basic aspects of standard biofeedback therapy should be considered. The basic pain biofeedback paradigm is a demonstration to the patient of the effect of internal emotional states on global physiological functions (eg, skin resistance, pulse, respiratory pattern, EEG) and focal processes (ie, electromyographic [EMG] activity) in regions of pain. Studies of several conditions have demonstrated the positive effects of this approach, usually combined with other physical therapy modalities. Peripheral pain
Headache
Generalized pain
VIRTUAL REALITY APPLICATIONS IN BIOFEEDBACKSection 5 of 10
VR techniques include a broad range of technologies. The common elements include an audiovisual system to present images and sounds controlled by the patient and his environment; a tactile input system, which can include hands; more extensive extremity contacts; and even whole-body environments. The VR system should allow for modification of stimuli based on the response of the patient. The degree of patient sensory involvement, or immersion, varies with system configuration.
An example of presented material would be a seascape. The images of ocean waves breaking with associated audio produce a soothing effect. The wave perspective and the volume of the sound can be made responsive to patient input, either directly (by manipulation of a glove or similar input device) or indirectly by measurement of skin resistance (GSR), EMG (cervical, facial, dorsal), pulse, blood pressure, and heart rate. EEG can be added as well. In a patient environment directed toward providing biofeedback, a totally immersive system carries increased risk of cybersickness and a feeling of isolation in some patients. In such patients, a graded approach can be used, with the initial system configured as an intermediate immersive system. Several scenes for relaxation and anxiety relief can be presented and switched by patient preference or biofeedback response. At the same time, a record is made of the patient's physiological response. Interaction with the therapist includes conventional biofeedback techniques. The power of this approach lies in the complete involvement of the patient in the biofeedback process. Traditional methods have employed counseling and patient observation of EMG, pulse rate, and other physiological functions. For many patients, this approach is anti-intuitive for two reasons. The graphic display of physiological functions often holds little interest for the patient, who is concerned primarily with symptom relief, not physiology. The interaction of the patient with the feedback environment primarily involves verbal communication with the therapist, who advises the patient on techniques of relaxation. While this patient-therapist interaction is important from an educational perspective, the core of the feedback process is involvement of the patient on a level of active and intuitive participation. The prolonged effects of biofeedback, when they are optimized, depend on the patient's awareness of the effects of altered perceptual and attentive states produced during the feedback process. Allowing the patient to realize that he can generate such alteration, both during the biofeedback exposure and independently, is the most important goal of therapy. Chronic pain produces an additive, integrative effect on patients. Not only the instantaneous severity of the pain, or anxiety, is deleterious. The cumulative effect of even moderate, but prolonged, pain can produce a higher overall level of discomfort and anxiety than pain of more severe intensity but shorter duration. A period of respite from the pain, and distraction from it, can reset pain thresholds and intensity perception levels. The same considerations apply to anxiety. Limbic interactions determining the level of discomfort evoked by pain are the basis for the music/white noise effect and the effects of hypnosis. The power of these effects, when properly harnessed, can be profound, and VR appears capable of providing a mechanism for achieving this result. Patients who have undergone cingulectomy report perception of pain, but without the negative affective component. Normal subjects can learn to dissociate the negative affective component of moderate, and even severe, pain from the sensation itself. VR techniques allow the patient to structure his environment to maximize this effect. By extension, many of the same considerations can be applied to free-floating and specifically localized anxiety states. VR environments have been used quite successfully in these settings. While conventional biofeedback techniques have generally been of little help in psychotic disorders, environment-based anxiety states have shown promising responses, with clinical improvement seen in fear of flying, acrophobia, and arachnophobia. The salient feature is creation of a sufficiently realistic environment to evoke the phobic response while maintaining patient cognition of the safety and control available in the test environment. As discussed below, by immersing the patient, while still maintaining patient control, the disparity between the phobic response and the actual danger becomes clear through the neuroplastic modification of the patient's responses and cue perception. These mechanisms parallel those involved in pain reduction. Neuroplasticity - Demonstration of potential The therapeutic promise of VR approaches in several neurological conditions is based on central neuroplasticity. Potential effectiveness has been demonstrated in motor disorders and in the cognitive and perceptual realms. This potential, both theoretical and clinical, for the expansion of these techniques into motor, sensory, and biofeedback interactive systems follows from a variety of laboratory and clinical situations. In addition to the standard biofeedback techniques, additional modalities are currently being developed, using custom-design video analysis techniques. Such a system includes the following components:
Clinical applicability of VR biofeedback Patients benefiting from this therapy initially would include those with subacute or chronic pain (eg, myofascitis, fibromyalgia, arthritis), posttraumatic pain, postsurgical pain, or headache (eg, migraine, tension, chronic daily, posttraumatic). As the program expands, application also may extend to certain psychiatric conditions, particularly anxiety states and phobic states. The system is likely to provide benefit to terminally ill patients. Another potential avenue for investigation is stress reduction in patients with multiple sclerosis. As an example of the power of the immersive technique, considerable reduction in pain is possible for burn patients during debridement by employing a video game or distracting virtual environment. The key elements in this approach are the immersion and the level of patient control of the virtual environment. Early reports indicate that a semi-immersive environment also may be helpful for patients with cancer. COMPUTER AS THERAPISTSection 6 of 10
In the previous section, Virtual Reality Applications in Biofeedback, the relationship between patient responses and the stimuli presented is emphasized. Traditionally, the therapist interprets physiologic responses, and the patient's demeanor, as confirmations of verbal descriptions from the patient regarding the efficacy of treatment. However, advances in technology have allowed more of that interpretive role to be transferred to the computer. VR biofeedback is a bidirectional interface comprising stimulus output modalities as described and input modalities that can receive not only basic physiologic responses but also changes in facial expression. Tracking of facial features has been available for several years on PC platforms, but what has not yet been fully developed is the extraction of behaviorally related data in scalar form for the processing video facial data. In the author's laboratory, active development has been proceeding (patent pending) in designing video-to-scalar conversion methods to produce statistically measurable correlation with behavior. For example, processing of video facial data yields eye blink, yawn, and head position measurements that are useful in determining states of alertness versus drowsiness. More precise measurements have been achieved with refinement of these techniques. Demonstration of these techniques is available through VR Neurotech. For example, data from the eye and mouth regions can produce highly specific characterization of emotional and arousal states. Image 1 shows this methodology applied to extract information from video monitoring of the face in the orbital region. The upper trace shows the position of the eyebrow region, and the lower trace shows the lid position (palpebral fissure). Good time resolution and quantitative position data are obtained and can be correlated with the patient's attention and emotional response. Similarly, Image 2 demonstrates data for the mouth region and includes good characterization of the time course of jaw positioning. This level of accuracy allows distinction between actual yawning and other types of mouthing movements, such as talking or even automatisms (for further discussion, particularly regarding partial complex seizure diagnosis, see Virtual Reality to Evaluate Motor Response During Seizure Activity). An example of such real-time acquisition is explained in Image 3. Here, both regions are acquired simultaneously. The colored rectangular areas superimposed on the raw video indicate the regions of interest (ROIs) where the image analysis takes place. The eye ROI activity is shown in the first and third traces, as labeled, and the mouth ROI activity is shown in the second and fourth traces, as labeled. This depiction explains the real-time acquisition shown in the movie clip in Image 3. Here, the actual acquisition process is seen in real time. Processing of this ROI configuration data allows the computer to interpret in real time whether the monitored movements are likely to represent physiologically significant activity. The movie clip in Image 4 demonstrates this process. The audio represents the real-time call-out by the computer of its interpretations: "open" for active eye opening, "close" for active eye closing, and "mouth" for movements that are candidates for relevant actions (yawning in this case). This behavioral filter represents a first approach based upon the techniques developed to date in the author's laboratory (patent pending). Further refinements are proceeding to improve the accuracy of these behavioral interpretive algorithms. Further studies are also in progress to relate these data as 12 channels obtained from 2 cameras for the prediction of drowsiness in drivers, as well as other behaviors relevant to the biofeedback sphere. By extension of the drowsiness work, changes in facial features will likely be of predictive value for determining anxiety versus relaxation. Such input to the biofeedback system described above would constitute a major functional improvement over collecting simple physiologic data, such as EMG and GSR, and would be noninvasive. A major obstacle to bringing these biofeedback techniques to general use by patients in the field is the complexity, with its associated cost, of currently available techniques. However, by using digital signal processing (DSP) methods, reducing the input devices for such patient monitoring data to stand-alone boards that can be produced quite inexpensively is possible, and these can thus be deployed widely. Such a DSP system is currently under construction in this author's laboratory. The computer responses need to be directed to produce an immersive environment. Optimally, this would comprise wide-angle, preferably 3D video, surround sound, posterior column stimulation (vibrational components), and, in selected environments, might even include transcutaneous nerve stimulation (in chronic pain situations). These responses need to be keyed and modulated by the patient's own responses. Biofeedback inputs, including facial feature changes, will control such responses. An example of such an approach is shown in Image 4 and in the movie clip in Image 6. Several types of stimuli, visual and auditory, can be optimized for individual patients and controlled by patient responses. For example, in the movie clip in Image 6, motion of the bird, the sea, and their intermixture can be modulated and correlated with the patient's response. The behavior of the artificial designs, including the rhythmicity of their movements and their interaction with the more natural stimuli, can also be controlled by the patient's response. In the example shown, sequences of a virtual pas-de-deux between the design avatar and the bird are present. The audio component intermixes music and sea sounds. Of course, because of technical limitations, the clip that can be shown on the Web encompasses only a fraction of the immersivity and quality of the stimuli that must be brought to bear on the problem. However, even in this clip, one can get a sense of the possibilities of the medium to tap the physiologic resources that exist in the human nervous system. Indeed, the patient can compose his or her own symphony of stimuli to interact with the intrinsic neuroplasticity involved in the generation of unpleasant sensations and their desired suppression. PSYCHIATRIC APPLICATIONSSection 7 of 10
The immersive environment has shown considerable promise as an intervention tool for several psychiatric disorders. The characteristics of immersion, attention redirection, cognitive resetting, and active environmental control by the patient all appear to contribute to its effectiveness. Following are a few examples of early indications of efficacy.
CONCLUSIONSSection 8 of 10
For excellent patient education resources, visit eMedicine's Muscle Disorders Center. Also, see eMedicine's patient education article, Chronic Pain. MULTIMEDIASection 9 of 10
REFERENCESSection 10 of 10
Virtual Reality Biofeedback in Chronic Pain and Psychiatry excerpt Article Last Updated: Dec 8, 2006 | ||||||||||||||||||||||||||||||||||||||||||
