Bilateral and Unilateral ECT: Effects on Verbal and Nonverbal Memory

By Larry R. Squire and Pamela Slater
American Journal of Psychiatry 135:11, November 1978

The memory loss associated with bilateral and nondominant unilateral ECT was assessed with verbal memory tests known to be sensitive to left temporal lobe dysfunction. Bilateral ECT markedly impaired delayed retention of verbal and nonverbal material. Right unilateral ECT impaired delayed retention of nonverbal material without measurably affecting retention of verbal material. Nonverbal memory was affected less by right unilateral ECT than by bilateral ECT. These findings, taken together with a consideration of the clinical efficacy of the two types of treatment, make what appears to be a conclusive case for unilateral over bilateral ECT.

Electroconvulsive therapy (ECT) has long been considered an effective treatment for depressive illness (1,2). The memory loss associated with this treatment has been well documented (3-5). For example, following conventional bilateral treatment, memory loss can extend to events that occurred many years before treatment as well as to events that occur during the weeks after treatment. Memory functions gradually improve as time passes after treatment. (6)

It has been generally accepted that right unilateral ECT is a clinically effective treatment that produces less impairment of new learning capacity and less amnesia for remote events than bilateral ECT (7-13). However, since right unilateral ECT is specifically associated with impairment in nonverbal memory (e.g., memory for spatial relationships, faces, designs and other material that is difficult to encode verbally (14-17), and since most studies of ECT and memory loss have employed verbal memory tests, the actual extent of memory loss associated with right unilateral ECT has remained somewhat unclear. It has been suggested that the amnesic effects of left or right unilateral ECT may be similar to the effects of left or right temporal lobe dysfunction (18). Accordingly, if memory were assessed with nonverbal tests specifically sensitive to right temporal lobe dysfunction, the amnesic effect of right unilateral ECT might prove to be as great as or even greater than that of bilateral ECT.

Only two studies have addressed this issue directly, employing verbal and nonverbal memory tests with patients receiving bilateral or right unilateral ECT. In the first study (15) impairment in one nonverbal test was somewhat greater after bilateral ECT than after unilateral ECT, but this difference was not statistically significant. In the second study (16) the results were ambiguous. Impairment in a nonverbal test was greater in the unilateral group after 4 treatments, but greater in the bilateral group 3 months after treatment. That study was further complicated by the fact that one-third of the patients given unilateral treatment did not have a grand mal seizure. Finally, since it was not clear how patients with identified right unilateral lesions would perform on the nonverbal tests used in these two studies, it was difficult to be sure how specifically sensitive the tests were to right hemispheric dysfunction.

The present study investigated memory functions in patients receiving bilateral or right unilateral ECT. Assessments of memory were made with two verbal tests known to be sensitive to left temporal lobe dysfunctions and two nonverbal tests known to be sensitive to right temporal lobe dysfunction.



The subjects were 72 psychiatric inpatients (53 women and 19 men) from 4 private hospitals, who had been prescribed a course of ECT. The diagnoses as recorded upon admission by the psychiatrists were depression (N=55); this diagnosis included designations of primary affective disorder, involutional melancholia, manic-depressive, and psychotic depression, neurotic depression (N=11), schizo-affective disorder (N=5), and hysterical personality (N=1). Patients with neurological disorders, schizophrenia with depression, depression secondary to alcoholism or drug abuse and patients who had received ECT during the previous 12 months were excluded from the study. Most of the patients (N=45) had not received ECT before; 27 had received ECT 1 to 15 years earlier.

The 72 patients in the study were assigned to 3 groups (table 1). Group 1 consisted of 33 patients who had been prescribed bilateral ECT. Group 2 consisted of 21 patients who had been prescribed right unilateral ECT. The choice of bilateral or unilateral ECT depended on the preferences of the individual psychiatrists and was therefore not random. However, since the patients about to receive bilateral or unilateral treatment did not differ measurably on their memory test scores before ECT (figure 1), it seems reasonable to assume that group differences emerging after ECT can be attributed to the type of ECT administered. Group 3, a control group, consisted of 18 randomly selected patients who were only tested before receiving a course of ECT. Fourteen of these patients were scheduled to receive bilateral ECT and 4 right unilateral ECT. All subjects were determined to be strongly right-handed; they reported that they did not use their left hand for any everyday activity and had no left-handed parent or sibling.


ECT was administered three times a week on alternate days following medication with atropine, methohexital sodium, and succinylcholine. Bilateral and unilateral treatments were administered using a Medcraft B-24 machine. For bilateral treatment electrode placement was temporal-parietal; for unilateral treatment both electrodes were placed on the right side of the head, as described by McAndrew and associates (19) (N=19) and by D'Elia (7) (N=10). Amnesic effects of nondominant unilateral ECT have been reported to be similar despite wide variation in electrode placement (20, 21). The stimulus parameters (140-170 v for .75-1.0 seconds) were sufficient to induce a grand mal seizure throughout the course of all treatments.

Tests and Procedures

Two memory tests, each consisting of a verbal and a nonverbal portion, were employed.

Test 1A (verbal portion: story recall). A short paragraph was read to the subject (6). Patients with identical dysfunction of the left temporal lobe are known to perform more poorly on this test than patients with dysfunction of the frontal parietal or right temporal region (22). Immediately after hearing the story, and again the next day (16-19 hours later), subjects were asked to recall as much as they could remember of it. The paragraph was divided into 20 segments, and the score was the number of segments recalled. Eighteen patients receiving bilateral ECT and 13 receiving right unilateral ECT were tested before treatment and again, with an equivalent form of the test, 6-10 hours after the fifth treatment of the series.

Test 1B (nonverbal portion: memory for geometric figure). Subjects copied a complex geometrical design (the Rey-Osterrieth figure [23] or the Taylor figure [24]) and were then asked to reproduce it from memory 16-19 hours later. Patients with right temporal lesions are known to be deficient on this task, whereas patients with left temporal lesions exhibit no impairment (25). The score for this test depended on the number of properly placed line segments (maximum score=36 points). The same patients given test 1A (above) were tested with one of these figures before ECT and with the other ones 6-10 hours after the fifth treatment.

Test 2A (verbal portion: short-term memory distractor test). Subjects were shown a consonant trigram, distracted for a variable interval (0, 3, 9 or 18 seconds), and then asked to recall the consonants (26). Patients with left temporal lesions are impaired on this task; patients with right temporal lesions are not (27). Subjects received 8 trials at each retention interval, and their score was the number of consonants correctly recalled without regard to order. The maximum score was 24. Fifteen patients receiving bilateral ECT were tested on two occasions with equivalent forms of this test. These sessions were scheduled 2-3 hours after the first treatment and 2-3 hours after the third treatment in the series. In addition, 8 patients receiving right unilateral ECT were tested 2-3 hours after their first and third treatments. Finally, 18 patients were tested on one occasion 1-2 days before their first treatment.

Test 2B (nonverbal portion: spatial memory). subjects attempted to remember the position of a small circle located along an 8-inch horizontal line. Patients with right temporal lesions are impaired on this task; patients with left temporal lesions are not (27). subjects inspected the circle on the line for 2 seconds and then were distracted for 6, 12 or 24 seconds by arranging strings of random digits into numerical order. Then subjects attempted to mark on a different 8-inch line the remembered position of the circle. Twenty-four trials were given, with 8 at each of the three retention intervals. The score on each trial was the distance (in millimeters) between the position of the originally presented circle and the position of the circle as marked by the subject. The score on the test at each retention interval was the total error (in millimeters) for all 8 trials. Test 2B was given on the same occasions and to the same patients as test 2A (above).


Figure 1 shows the results with test 1 for patients who received bilateral or unilateral ECT. Before ECT these two groups of patients did not differ from each other on any of the measures of immediate or delayed recall (for the verbal test t<1.5, p>.10; for the nonverbal test, t=0.7, p>.10). After ECT patients receiving bilateral treatment were able to remember verbal material immediately after hearing it as well as they could before ECT (before ECT versus after ECT, t=0.1, p>.10), and they were able to copy a complex figure as well as before ECT (t=0.1, p>.10). However, their performance was severely impaired on delayed tests of verbal and nonverbal memory (verbal test: before ECT versus after ECT, t=5.6, p<0,1; nonverbal test: before ECT versus after ECT, t=3.7, p<0.1).

Right unilateral ECT did not affect verbal memory, as measured by test 1A. That is, the delayed recall scores of patients receiving right unilateral treatment were about the same after ECT as before (t=0.6, p>.10). However, nonverbal memory was significantly impaired by right unilateral ECT (test 1B). Before unilateral ECT the score for reproducing the geometric figure after a delay was 11.9, and after unilateral ECT the corresponding score was 7.1 (t=2.7, p<.05). This impairment in nonverbal memory associated with unilateral ECT was not as great as the impairment in nonverbal memory associated with bilateral ECT (t=2.1, p<.05).

Figure 2 shows the results with test 2 for patients receiving bilateral ECT, patients receiving right unilateral ECT, and a control group of patients about to begin a course of bilateral or unilateral ECT. For the short-term memory distractor test, patients receiving bilateral ECT were impaired, but patients receiving right unilateral ECT performed normally. An analysis of variance with repeated measure on one factor (28) indicated that the scores of bilateral patients were significantly lower than those of both unilateral patients (F=10.8, p<.01) and control patients (F=5.7, p<.01). The scores of unilateral patients and control patients were not measurably different (F=0.8, p>,10).

For the spatial memory test bilateral ECT also produced a marked impairment (bilateral group versus control group, F=22.4, p<.01). The scores of unilateral patients were also poorer than those of control patients, although this difference fell short of significance (F=2.64, p=.12). Finally, the effect on nonverbal memory associated with unilateral ECT was not as great as the effect associated with bilateral ECT (F=9.6, p<.01).


The results can be summarized by three main conclusions.

1. Bilateral ECT markedly impaired the ability to retain both verbal and nonverbal material.
2. Right unilateral ECT impaired the ability to retain nonverbal material without measurably affecting memory for verbal material.
3. The impairment in nonverbal memory associated with right unilateral ECT was less than the impairment in nonverbal memory associated with bilateral ECT.

The findings that bilateral ECT markedly affected memory and that right unilateral ECT exerted a material-specific effect on nonverbal memory are consistent with the results of a number of studies of ECT and memory loss (3-5, 7). However, it should be noted that the extent to which bilateral or right unilateral ECT impairs memory depends on the sensitivity of memory tests to the effects of ECT. For example, in the present study right unilateral ECT had no measurable effect on verbal memory; yet performance on some verbal memory tests can be impaired by right unilateral treatment (10,12). Accordingly, it is difficult to compare the amnesic effects of bilateral and right unilateral ECT unless these effects are assessed in the same study using the same tests.

The present study employed memory tests known to be sensitive to either left or right temporal lobe dysfunction. The results clearly indicated that the effect of right unilateral ECT on both verbal and nonverbal memory was less than that of bilateral ECT. It has sometimes been assumed that right unilateral ECT produces as much memory dysfunction as bilateral ECT on those aspects of memory function associated with the right hemisphere. To our knowledge, the study reported here is the first to clearly demonstrate that right unilateral ECT produces less memory dysfunction for nonverbal material than bilateral ECT.

The therapeutic efficacy of bilateral and unilateral ECT has been compared in a large number of studies (for reviews see references 29 and 30). Taken together, these studies indicate that courses of bilateral or unilateral ECT are approximately equivalent. They lead to similar reductions in depressive symptoms, are associated with similar relapse rates, and exhibit similar efficacy at follow-up. One review (29) has suggested that the slight disadvantage in immediate efficacy sometimes reported for unilateral treatment, as well as the apparently widespread impression (footnote 1) that unilateral ECT is not as effective as bilateral ECT, may be due to occasional failures to produce a maximal seizure with the unilateral technique. Since the therapeutic effect of ECT is bound to the seizure (32), even one sub-maximal seizure during a course of unilateral treatment could account for reported slight differences between unilateral and bilateral ECT. Several practical suggestions to ensure that unilateral ECT produces a grand mal seizure have been outlined (29).

When given properly, unilateral ECT seems to be clearly preferable to bilateral ECT since the risks to verbal and nonverbal memory are less than for bilateral treatment. It should be noted that some risks to memory exist even for unilateral ECT. The benefits to be derived from this procedure should therefore be weighed carefully against these risks and against the possible risks of alternative therapies to form a basis for clinical judgment.

1. A recent survey of members of the American Psychiatric Association conducted by the APA Task Force on ECT indicated that of 3,000 respondents, 75% of those who used ECT used bilateral for all their patients. (31)