Physiological measures were taken from 7 closed-head-injured patients and 7 control subjects while they took a series of cognitive tests: the finger tapping test from the Halstead Reitan battery, the Digit Symbol test from the WAIS-R, a test of logical memory, and a paced arithmetic test. Physiological reactivity was assessed relative to rest periods, which occurred at the beginning and end of each session. The tests and physiological assessment were administered twice, approximately one month apart. The patient group performed more poorly on the cognitive tests, and showed less physiological reactivity during them than did the control group. This pattern was statistically significant for heart rate, frequency of electrodermal responses, and, during the initial session, respiration rate. The control group also showed greater finger pulse amplitude during the first posttest rest period in the first session, and greater constriction during a logical memory task in the second session. No significant between-groups differences emerged for state or trait anxiety. The patients showed higher frontalis EMG and respiration rate during rest. These results suggest a pattern of poor physiological modulation for task performance in the patients with closed head injuries. The therapeutic implications are discussed.