An initial flaccidity usually follows a motor cortical capsular lesion. This gives way to an emerging spasticity which is most com- monly predominant in the flexors of the arm and extensors of the leg.

Clonus is a pathological oscillation which may be myoelectrically exhibited by the muscle. It is elicited if the muscle is passively stretched and a tendon stretch reflex is initiated. After the initial tendon stretch reflex contraction, the muscle relaxes and stretches the muscle spindles, causing a synchronous resumption of spindle afferent discharge. The synchronous increased monosynaptic projections of the spindle afferent discharge cause the homonomous alpha motor neurons to discharge again, producing a second stretch reflex contraction of the muscle. As the muscle subsequently relaxes, the cycle is repeated. Clonus occurs because the chronictonic shortening of the muscle spindle predisposes hyperreflexive stretch reflexes.

In spastic muscles, the hyperexcitability of the alpha or gamma motor neurons are such that the stretch reflex and subsequent synchro-nous muscle spindle firing begin to cycle discharges of the alpha motor neurons and stretch reflex contractions. In cases of severe spasti- city, clonus commences if the muscle is inadvertently stretched, and may persist for extended periods until the initiating stretch is stop- ped. Such a closed, oscillating feedback loop indicates the importance of inhibitory control mechanisms which normally promote assynchrony of neural activity.

Review of Literature:
Marinacci and Horande (1960) investigated left-sided hemiplegia. They inserted EMG needle electrodes into the involved left arm muscles, and could find no voluntary nerve impulses. They then inserted electrodes into the normal right deltoid and showed the patient how muscle activity could produce auditory feedback. The electrodes were then inserted into the paralyzed left deltoid muscle. The patient was able to generate from 10 to 15 percent of motor units in a location from which there had been no previous detectable activity. The same procedure was utilized successfully at other muscle sites. This study represents the initial pioneer EMG study for hemiplegia.

Andrews (1964) reported on an uncontrolled study utilizing a patient group of hemiplegics who had EMG electrodes inserted in the involved triceps muscle. Auditory feedback was provided as the subject tried to generate sound and movement. A five-minute trial period was allowed, and 17 out of the 20 patients showed an increase in motor action potentials.

Johnson and Garton (1973) reported on ten hemiplegic patients, who utilized EMG as an aid in total rehabilitation rather than just the re-turn of voluntary movement as in the Andrews study. Five out of ten subjects had enough improvement to eliminate leg bracing on the in-volved side. It was a single group outcome study without a control group but, the group did have a prolonged baseline period of one year without functional return before treatment.

Brudny, et. ad. (1974) used EMG feedback to treat a group of 36 patients, 13 of whom had hemiparesis. In this study, surface elec- trodes were used instead of inserted needle electrodes. In two indi-viduals, there was no change. In one patient, there was relief from muscle spasticity. In six patients, function of the extremity was re-established, and in four cases, prehension became possible. These were anecdotal case reports with long baselines existing before treatment.

Swaan, Van Wieringer and Fokkema (1974) explored EMG feedback of seven patients, four of whom were hemiplegic. They were taught to inhibit the peroneus longus muscle while contracting the quadriceps muscle. Conventional rehabilitation methods were used to suppress the undesirable hyperactivity of the peroneus longus muscle along with the feedback. No justification was given for reinforcement of the quadri-ceps and inhibition of the peroneus longus.

Basmajian, Kakulka, Narayan, and Takebe (1975) compared EMG bio- feedback plus physical therapy with the results of standard rehabili-tation procedures in cases of ankle dorsiflexion paralysis after stroke. The authors claimed that an increase in both strength and range of motion in the biofeedback group was twice as great as the achievements of the exercise control group. The two groups of patients were not variably matched. When biofeedback was added to physical therapy, the mixed variables were not controlled.

Taylor and Bongar (1976) described the use of electromyometry feedback for the treatment of cerebrovascular lesion patients. Patients were taught to inhibit one set of muscles while simultaneously facilitating others. For example, inhibition of thumb flexion was attempted while thumb extension was facilitated. Their approach was anecdotal. The methodology has subsequently been expanded upon and is described in the proceeding chapters.

Keefe and Trombly (1979) utilized EMG feedback to aid a hemiplegic patient judge limb position without being able to see the limb. The patient participated in an A-B-A-B withdrawal design to evaluate the effects of EMG biofeedback on accurate limb positioning. EMG feedback with audio feedback produced a decrement in performance, and when EMG feedback was re-instituted, performance once again improved. The patient was able to generalize the EMG feedback training to improve functional use of the arm.

Koheil, et. al. (1979), at the Ontario Crippled Children's Center, developed a Joint Position Trainer to provide precise feedback of limb position to three hemiplegics. The Joint Position Trainer provided feedback of position rather than of muscle activity, and incorporated a goniometer attached to a leg cuff with auditory feedback of knee joint angle. Two of the three patients developed improved gait with increased control of knee hyperextension.

The Study:
Subjects: The present study was designed in 1978, as a within- group-controlled outcome study design for six CVA patients, all of whom were at least three years post stroke. Three were aphasic, and three were not.

Method: Myoelectric activity from the finger flexors and finger ex- tensors was monitored with EMM feedback as the muscles were balanced against one another during neuromuscular re-education. The subjects' hands were stabilized by being taped to a board with the fingers ex-tended, and the thumbs extended and abducted. Subjects picked for the study were considered unlikely to improve, as they were more than three years post injury, and none had dominant hand extension over hand flex-ion. The total study time was 15 months. The best EMM readings from each muscle group during each half-hour session were collected by the physical therapy staff. These readings were taken each biweekly ses-sion for ten months. At the end of the ten months, each individual in the study received an additional 30 minutes of electroencephalographic feedback on the left sensorimotor cortex, T3C3, according to the Inter-national 10- 20 Electrode Placement System. The EEG feedback training continued biweekly for five months. Each EMM feedback training session occurred just after each EEG training session. EMM recordings were collected as before. The EMM instrumentation used was the Toomim elec-tromyometer, and the EEG was the Neuroanalyzer 400, designed by Sidney Ross (see Illustration 66). A picture of the International 10-20 Elec-trode Placement System is presented in Figure 23.

Rationale: Electroencephalographic feedback from T3C3, sensorimotor cortex, was utilized because changes in activity from this area have been correlated with muscle activity in the hand by Kato and Tanji (1972). They reported EEG changes in respect to the contraction of a whole muscle, specifically in the abductor pollicis brevis and the opponens pollicis muscles. Needle electrodes were used for both EEG and EMG, and the EEG was recorded by a needle electrode inserted sub-cutaneously into the scalp contralateral to the finger studied. Elec-trical activities from other ipsilateral muscles, such as the brachio-radialis, biceps, deltoideus, trapezius, and sternocleidomastoideus were checked but, no simultaneous discharges in those muscles occurred. EEG changes that accompanied the singled-out motor unit discharges were largest when recorded from the contralateral finger area than from any other site in the scalp.

Instrumentation and Method:
In this research study, the Neuroanalyzer 4000 included in its' construction tuned active filters, an analog analysis circuit and a timing circuit. Subjects were attached with bipolar leads to T3C3, directly over the sensorimotor cortex, with gold plated Grass cup electrodes, applied with Grass EEG paste. The reference ground was placed on the ear lobe at A1. A strip recorder displayed the raw EEG, which included filtered frequency analysis of 12 to 15 hertz, 15 to 18 hertz, and 4 to 7 hertz.

The Neuroanalyzer 4000 processes EEG signals from the patient to produce visual and auditory feedback as a means of retraining brainwave patterns. EEG electrodes on the individual are connected to the Neuro-analyzer, and the EEG signal is amplified to a high level signal. Four separate channels in the analyzer process this high level EEG signal. Each channel analyzes the specific frequency band present in the EEG signal. Other channels are singled out for frequency inhibition. The latter two are labeled Inhibit B and Inhibit C. One channel is re-served for a frequency to be facilitated while two others are inhi-bited, B and C. Inhibits B and C are tuned to two different frequency bands. Inhibit B is the primary frequency inhibited, and Inhibit C is the highest frequency, 20 to 25 hertz (cycles per second), which makes a ceiling level of narrowed range within which the patient can work. The last channel, Inhibit A, detects any undesired high amplitude sig-nals present in the EEG signal, such as body movement or artifact. When a specified frequency is produced at the desired amplitude for a half-second, a visual and auditory reward is produced but, only if the undesired frequencies are inhibited sufficiently for half a minute. The number of rewards produced are totaled on an electronic digital counter. When any of the undesired signals are detected in a suffi-cient amount, one of two red inhibit lamps will illuminate, and the reward signals will be inhibited. The amount (amplitude sustained for a half-second) of the desired frequency is displayed on a variable intensity display. As the amount of the rewarded frequency is in-creased, the display glows brighter, and as it decreases, the display dims. The rewarded display is inhibited if either of the inhibit channels is activated. On the front panel, controls can be changed to adjust the voltage levels for the four channels. The resulting raw EEG, filtered frequencies and relays displaying amount of frequency are recorded on EEG polygraph paper.

Since CVA lesions have been characterized by high amplitude slow waves, and even K complexes that have been evoked by somatic stimula- tion (according to Goff, et. al., 1971), it was decided to have sub- jects inhibit 4 to 7 cycles per second, while producing or being re- warded for 12 to 15 cycles per second. K complexes have been associ- ated with the production of the sleep spindle in stage two of sleep, which is 12 to 15 cycles per second.

The rationale for selection of the 12 to 15 cycles per second setting was based on the noted association by Chase and Harper (1971) between the production of 12 to 15 hertz and behavioral immobility and the application by Sterman and MacDonald (1978) of central cortical EEG feedback training in poorly controlled grand mal epileptics with re-enforcing production of 12 to 15 cycles per second resulting in a re-duction of seizures. However, neither Chase and Harper nor Sterman and MacDonald postulated the possibility of proprioceptive sensory inhibi-tion resulting in the increased behavioral immobility or of a reduction in seizure motor discharge.

For statistical purposes, the finger extensor and finger flexor activities in microvolts were combined into a ratio of extensor acti-vity divided by flexor activity. This ratio was recorded for the full 15 months of EMM record taking.

Hypothesis: There were two hypotheses:

1. The EEG feedback intervention would have a significant in-crease upon finger muscle performance at the .05 level of significance.


2. The variables of age, side of paresis, number of years post stroke, amount of skilled hand use, and the sex of the sub-ject would not affect EMG performance prior to or during EEG intervention.

Results:
Each subject's mean Emm activity ratio was calculated for the ten months prior to EEG intervention (the pre-EEG group), and for the five months after EEG intervention (the post EEG group). The paired t statistic was calculated from the differences between these mean scores. The first hypothesis was accepted. EEG intervention had a significant effect of increasing EMM performance (p .025, T = 2.4949, df = 6). Examples of the EEG of one cerebrovascular lesion patient in the study prior to EEG intervention and post EEG intervention may be found in Figures 24 and 25. Figure 26 provides a representative sample of the patient's EMM/EEG performance throughout the 15-month study period.

For hypothesis two, the variables were submitted to the multiple regression procedure with EMM score as the criterion variable. The regression results indicated two significant factors in the equation. These were the left side of paresis (F = 27.893, p .001, df = 5.123) and the age of the subject (F = 21.168, p .001, df = 5.123). The other variables were not significant, that is, right side of paresis, number of years post stroke, amount of hand use, and the sex of the subject.

Discussion:
Of particular interest was the high level of significant for both hypotheses being accepted. In addition, individuals did not improve in gaining dominance of finger extension over finger flexion during previous EMM training but, changed almost immediately when EEG feed- back was initiated.

Possibly, the failure to gain finger extension during initial EMM feedback could have been a function of the severity of the stroke, and the dominance of abnormal reflex patterns. The rapid gain of extension with EEG intervention could be due to combining peripheral EMM feedback with central cortical EEG feedback. Those individuals with a left-sided paresis had greater gains in finger extension than those indivi-duals with a right-sided paresis. The three individuals with a lesion on the left and three individuals with a lesion on the right were all trained on left sensorimotor cortex. The older the patient, the more gains were obtained in finger extension.

Never before has EEG feedback been utilized with cerebrovascular lesion patients, nor has there been a controlled research study utiliz-ing EMM and EEG feedback until now. Questions should be raised as to the role of afferent input, particularly proprioceptive, in relation to motor activation and inhibition of sensorimotor cortex in utilizing EEG and EMM feedback.

"A Report on a Study of the Utilization of Electro-encephalography (Neuroanalyzer) for the Treatment of Cerebral Vascular Lesion Syndromes", Chapter 7 in Electromyometric Biofeedback Therapy by Taylor, L.P.; Ayers, M.E.; and Tom, G.; Biofeedback and Advanced Therapy Institute, Los Angeles, California, pages 244-257, 1981.

Andrews, J.M., "Neuromuscular Re-Education of the Hemiplegic With the Aid of the Electromyograph", Archives of Physical Medicine and Rehab-ilitation, 45:530-532, 1964.

Basmajian, J.V., Kukula, C.G., Narayan, M.G., and Takebe, K., "Biofeed-back Treatment of a Foot-Drop After Stroke Compared With Standard Re-habilitation Techniques: Effects on Voluntary Control and Strength", Archives of Physical Medicine and Rehabilitation, 56:231-236, 1975.

Bates, J.A.V., "Electrical Activity of the Cortex Accompanying Move- ment", Journal of Physiology, (London), 113:240-257, 1951.

Brudny, J., Korein, J., Levidow, L., Grynbaum, B.B., Lieberman, A., and Friedman, L.W., "Sensory Feedback Therapy as a Modality of Treatment in Central Nervous Disorders of Voluntary Movement", Neurology, 24:925-932, 1974.

Chase, M.H., and Harper, R.M., "Somatomotor and Visceromotor Correlates of Operantly Conditioned 12-14 Cycles Per Second Sensorimotor Cortical Activity", Electroencephalography and Clinical Neurophysiology, 31:85-92, 1971.

Goff, W.R., Bobowick, A.R., Allison, T., and Levy, L., "K Complexes Evoked By Somatic Stimulation in Patients with Unilateral Cerebral Lesions", Electroencephalography and Clinical Neurophysiology, 31:289, 1971.

Johnson, H.E., and Garton, W.H., "Muscle Re-Education in Hemiplegia By Use of Electromyographic Device", Archives of Physical Medicine and Rehabilitation, 54:320-325, 1973.

Kato M., and Tanji, J., "Cortical Motor Potentials Accompanying Voli- tionally Controlled Single Motor Unit Discharges in Human Finger Muscles", Brain Research, 47:103-111, 1972.

Keefe, F., and Trombly, K., "Impaired Kinesthetic Sensation: Can EMG Feedback Help?" Presented at the Proceedings of Biofeedback Society of America, Tenth Annual Meeting, February of 1979, in San Diego, California.

Koheil, R., Mandel, A., Herman, A., and Iles, G., "Joint Position Training For Hyperextension of the Knee in Stroke Patients: Prelimi- nary Results", Presented at the Proceedings of the Biofeedback Society of America, Tenth Annual Meeting, February of 1979, in San Diego, California.

Marinacci, A.A., and Horande, M., "Electromyogram in Neuromuscular Re- Education", Bulletin of the Los Angeles Neurologic Society, 25:57-71, 1960.

Sterman, M.B., and MacDonald, L.R., "Effects of Central Cortical EEG Feedback Training on Incidence of poorly Controlled Seizures", Epilep- sia, 19:207-222, 1978.

Swaan, D., Van Wieringen, P.C.W., and Fokkema, S.D., "Auditory Elec- tromyographic Feedback Therapy to Inhibit Undesired Motor Activity", Archives of Physical Medicine and Rehabilitation, 57:9-11, 1974.

Taylor, L.P., and Bongar, B., Clinical Applications in Biofeedback Therapy, Psychology Press, Los Angeles, California, 1976.