Psychiatric Annals


ECT: Lessons Learned About an Old Treatment with New Technologies

John L Black, MD


Despite misgivings, ECT is a highly effective treatment for refractory depression, severe mania, some acute psychotic illnesses, and catatonic conditions.


Despite misgivings, ECT is a highly effective treatment for refractory depression, severe mania, some acute psychotic illnesses, and catatonic conditions.

Electroconvulsive therapy (ECT) is one of the oldest somatic treatments used for menial illness, but it remains one of the most controversial. Next to psychosurgery, electroconvulsive therapy is least accepted by physicians and nonphysicians. Despite these misgivings, ECT is a highly effective treatment for refractory depression, severe mania, some acute psychotic illnesses, and catatonic conditions. Furthermore, ECT is gaining acceptance for the treatment of refractory parkinsonism and neuroleptic malignant syndrome.

Newer technologies and, in particular, advanced imaging techniques are beginning to provide insight into how ECT affects the brain, and may eventually show us why it is effective. The advanced imaging modalities that will be discussed in this article include computed tomography (CT), magnetic resonance imaging (MRI), p 133Xenon (p 133Xe) inhalation, single photon emission computed tomography (SPECT), and positron emission tomography (PET). These new technologies permit us to examine the questions, the controversies, and the scientific basis for this treatment.

This article addresses three groups of questions:

* What is the effect of ECT on brain structure? Does ECT lead to brain damage?

* Is there an association between CT and MRI abnormalities and responsiveness to ECT or side effects from ECT?

* What effect does ECT have on brain function? Does this help us to understand why patients become depressed and how ECT works?


Friedberg1 stated in 1977 that electroconvulsive therapy caused brain damage. However, even then pneumoencephalograms (PEG) performed by four researchers revealed no evidence that those treated with ECT showed morphologic differences from those who were not treated. Another study showed an association between ECT treatment and abnormal PEGs. The authors stated that the sickest patients were likely to be sent to ECT, and therefore, this association did not prove ECT caused brain damage. They correctly assumed that these severely ill patients had abnormal CNS morphology before ECT, due to their psychiatric illness or other factors.2

With the development of advanced imaging technologies such as CT and MRI, the ability to image living CNS structure improved. Several findings must be remembered when reviewing data produced by these technologies. First, schizophrenics have shown increased ventricular- to-brain ratio (VBR), increased sulcal widening, and frontal atrophy. Likewise, major depressive illness and bipolar affective disorder have been associated with enlargement of the lateral ventricles or increased VBR.3-4 Certain findings have also been associated with aging. These include white matter changes known as periventricular hyperintensities (PVH), deep white matter hyperintensities (DWMH). leukoencephalopathy. or cerebromalacia. These lesions are believed to be associated with increased water content in the tissue due to cerebrovascular disease.5

With these findings in mind, one can more meaningfully understand the studies that have examined the effects of ECT on CNS structure. Many of the early studies were done without the knowledge of the association between CNS changes, age, and mental illness. Thus, some authors concluded that ECT caused the CNS changes noted.

CT Studies

Of the early studies done using CT, three retrospective studies were interpreted to show evidence that ECT affected CNS morphology.6 Weinberger7 found an increased VBR in ECT-treated schizophrenic patients and noted cortical atrophy more commonly in schizophrenic patients treated with ECT. Calloway and colleagues8 studied 41 elderly depressed patients and found an association between ECT and global cortical atrophy.

Several studies that included control groups yielded different results. Kolbeinsson et al6 studied 22 patients with a mixture of affective disorders who had received ECT and compared them with age- and sex-matched controls. No association was found between VBR and cortical atrophy scores and ECT.

Andreasen et al3 measured the VBR in 24 bipolar patients, 27 unipolar patients with major depression, 108 schizophrenic patients, and 75 normal control subjects. The male bipolar patients had increased VBRs, but this did not correlate with a history of previous ECT treatments.

Nasrallah9 studied 24 manic males and identified large and small VBR groups. Neither group was associated with a history of previous ECT.

Owens et al2 studied 112 institutionalized chronic schizophrenic patients and compared them with a group of noninstitutionalized mentally ill patients. Ventricular enlargement was unrelated to past physical treatments including ECT.

Prospective studies provide the most convincing evidence that ECT does not cause morphologic changes. Bergsholm et al10 performed CT before and within 24 to 40 weeks after unilateral ECT. Only one patient had questionable dilation of the left temporal horn, but the authors did not believe that this was due to neuronal death caused by ECT. The patient's changes on CT scan were felt to be due to normal aging, late onset depression, evolving dementing disorder and/or hypertension, rather than ECT.

MRI Studies

MRI is an easier technology to use than CT because radiation is not involved and prospective studies carry less risk. MRI is useful in examining anatomic changes and changes in the blood-brain barrier leading to fluid shifts. Mander11 measured the T1 relaxation time in 14 patients with major depressive illness receiving ECT. The T, relaxation time rose immediately after the seizure and reached a maximum in four to six hours. The T1 value then returned to the original level with no long-term increase occurring over the course of a series of treatments. The authors interpreted these results as indicating a temporary breakdown of the blood-brain barrier.

Scott12 obtained similar results with 20 unipolar depressed patients who were scanned before ECT treatment and at 25 minutes, 2 hours, 4 hours, 6 hours, and 24 hours. There was an immediate increase in T1 values peaking at about 1% higher within two hours. These changes gradually returned to baseline values. The authors found no correlation between MRI changes and the time patients took to become reoriented after ECT.12

Structural studies of ECT-treated patients using MRI reflect the findings with CT as reviewed by Coffey.5 However, MRI shows PVH and DWMH, which can be studied.

Coffey4 found a relatively high incidence of cortical atrophy, white matter changes (PVH, DWMH), and lateral ventricular enlargement in elderly patients with depression referred for ECT.

Pande13 studied 10 patients referred for ECT and got essentially the same results. These authors noticed that white matter lesions remained unchanged with electroconvulsive therapy.

Coffey4 prospectively examined brain structure changes in 35 inpatients with depression who underwent ECT for treatment of that depression. He measured regional brain volumes and performed pairwise global comparisons. He found no acute or delayed (six month) change in brain structure as measured by alterations of the total volume of the lateral ventricles, third ventricle, frontal lobes, temporal lobes, or amygdalahippocampal complex. Five subjects did show increased subcortical hyperintensity, which was believed to be secondary to progression of ongoing cerebrovascular disease.5

In summary, CT and MRI studies do not indicate that ECT causes permanent morphologic changes in the brain. It is interesting to note that many of the patients who received ECT had MRI abnormalities prior to EOT In individuals with altered anatomy, no changes occurred with ECT.

With regard to T1 changes noted by several authors, it is useful to categorize relaxation time changes as hyperacute, acute, and chronic. In the hyperacute state (less than one hour), T1 relaxation values increase. However, this usually peaks within the acute time frame (within two hours) and then gradually returns to normal. There is no cumulative effect recorded. The T1 changes have not been associated with disorientation or agitation.

It is possible that the changes in T1 values are due to breakdown of the bloodbrain barrier. It has been hypothesized that ECT acts by disrupting the bloodbrain barrier, which allows neuronally active peptides to have an effect on the activity of the central nervous system, thus alleviating psychiatric symptoms.


Coffey et al5 used MRI and CT to identify changes in subcortical white matter in 44 (66%) of 67 elderly depressed inpatients referred for ECT. Of these patients, 58% had developed late onset depressive disorders, and 68% had been refractory to or intolerant of antidepressant drug therapy. All but one of the 44 patients subsequently responded to a course of ECT. Six of the patients ( 14% ) developed interictal confusion. The authors note that the long-term prognosis of depression in patients with white matter changes has not been systematically studied. However, they note that brain imaging studies may provide important prognostic information for at least some elderly patients with depression. This leads to speculation that refractory depression in elderly patients may be an organic illness caused by the CNS changes noted in the study.

Sandyc et al14 examined the relationship between ECT nonresponsiveness and calcification of the pineal gland in bipolar patients as assessed by CT. They found an association between ECT nonresponsiveness in the presence of pathologically large pineal calcifications in 17 chronically institutionalized bipolar patients. The authors quote evidence that there is an inverse relationship between melatonin secretion and the degree of pineal calcification, suggesting that antidepressant effects of ECT could be mediated in part through activation of pineal melatonin secretion. These studies need to be reproduced using larger cohorts and control groups. However, one might speculate that MRI or CT imaging of the brain may help predict which patients will be refractory to psychopharmacotherapy (those with leukencephalopathy ) and ECT (those with large pineal calcifications).

In three studies, Figiel et al15 attempted to determine whether abnormalities seen on MRI prior to ECT predispose a patient to interictal confusion and delirium. In the first study, patients were examined for pre-ECT MRI abnormalities of the brain, including cortical atrophy, white matter lesions, lateral ventricular enlargement, and lesions of the pons. He noted an apparent association between white matter lesions and late onset depression. The severity of these lesions seemed to have some effect on reorientation times; those patients with more severe lesions tended to require more time to reorient. However, this finding was not statistically significant. Furthermore, on memory testing, there was no association between white matter lesions and difficulties with recall upon testing. The most remarkable finding was that four (80%) of five patients who developed interictal delirium had lesions of the basal ganglion. Twenty (43%) of 47 patients who did not develop delirium had basal ganglion lesions. Of 27 patients without lesions in the basal ganglion, only one (4%) patient developed an interictal delirium.

In the second study, Figiel noted that a prolonged delirium was induced by ECT in six (17%) of 36 elderly patients. Brain MRI or CT revealed structural changes in the basal ganglion and white matter of all six patients who developed delirium. In three of the patients, the right caudate nucleus had hyperintensities, and in one patient, both the right and left caudate nuclei had hyperintensities on MRI scans. In the two patients who had CT scans, each had hypodense areas in the left caudate nucleus. Unfortunately, brain scans were not obtained in the 30 patients who did not develop delirium. Thus, the frequency of caudate or white matter structural changes in that group could not be compared with the delirious group.

Finally, in the third study, Figiel prospectively studied 87 elderly depressed patients. Ten (11%) of the studied patients developed a prolonged interictal delirium during a course of ECT Of these, nine (90%) had basal ganglion lesions compared with 30 (39%) of the patients who did not develop interictal delirium. In addition, nine (90%) of the patients who developed interictal delirium had PVH and DWMH that were rated as moderate to severe. The groups did not differ with regard to moderate-to-severe cortical atrophy.

Figiel notes that literature on stroke patients supports the concept that individuals with basal ganglion lesions, and in particular, right basal ganglion lesions, are more susceptible to agitated, confused, delirious behavior. Also, patients with basal ganglion and subcortical white matter lesions may be more susceptible to antidepressant-induced delirium. The exact etiology of this delirium is unclear, but the basal ganglion and subcortical white matter have extensive connections with cortical areas known to be important to attentional processes. With lesions of the basal ganglion and subcortical white matter, it is likely that there is disruption or disconnection of these pathways resulting in increased vulnerability to delirium. If these studies can be reproduced, it may be reasonable to mitigate the degree of delirium in patients treated with ECT by first scanning the patients and then reducing the frequency with which they receive their ECT treatments.

In summary, with further study, CT and MRI promise to help determine:

* which patients are prone to refractory depressions that respond to ECT (e.g., those with white matter changes);

* which patients are not likely to respond to ECT (e.g., bipolar patients with calcification of the pineal gland); and

* which patients are likely to develop delirium when receiving ECT (e.g., patients with caudate nucleus and, in particular, right caudate nucleus lesions).


CNS function can be studied with several new imaging modalities. Regional cerebral blood flow (RCBF) and hemispheric cerebral blood flow (HCBF) are measured by using 133Xe inhalation and by SPECT using Tc99M-hexamethylpropyleneamineoxim (HMPAO) or N-isopropyl-p-(123I) iodoamphetamine (IMP). RCBF is considered to be an indirect measure of metabolic activity. PET can be used for RCBF studies by employing a variety of agents including 15O water (H2 15O) but [18F] fluorodeoxyglucose (FDG) studies are a more direct measure of the metabolic activity of neuronal tissues, and some researchers consider them better.

To understand the effect mental illness has on these studies, normative values must be established. Depression has been most carefully studied with these advanced technologies. Depressives studied with 133Xe inhalation techniques show a slight, increase in HCBF or RCBF.16 Paradoxically, in major depressive episodes, IMP SPECT shows a global decrease in RCBF over the cerebral hemispheres, with left being more decreased than right.17 PET studies of depressed patients using FDG obtained results similar to the SPECT studies.

Hurwitz18 found a significant absolute reduction in frontal metabolism most apparent in the anterior and right frontal cortex of depressed patients when compared with normal controls. Cortical hypometabolism was lateralized, being more prevalent on the right and in the dominant hemisphere. Other authors have found left dorsal anterolateral hypometabolism,19 whole brain hypometabolism,211 and lower caudate-tohemi sphere ratio.21 Thus, the true nature of the CBF and the metabolic rate of the central nervous system in major depressive episodes is unclear, making it difficult to interpret uncontrolled studies that examine the effects of ECT on these parameters. However, there is a trend toward these studies showing reduced metabolic activity in the prefrontal regions.

The effect ofECT on RCBF has been described in a series of articles by Silfverskiold22 using 133Xe inhalation techniques correlated with EEG results. During the ictal phase, CBF shows a large increase reflecting the increased neuronal activity. In both unilaterally and bilaterally administered ECT, slowing of the EEG and reduced CBF occurs immediately postictallv with the first ECT treatment. There is an increased slowing of EEG activity during serial ECT. EEG changes associated with a series of ECT reversed slowly upon completion of the series. Changes in CBF did not correlate with changes in EEG.

SPECT measures of RCBF reflect those seen with 133Xe. Baje et al23 measured RCBF with HMPAO SPECT during ECT-induced ictus and found a relative decrease in flow to the occipital cortex and thalamus but increased flow in the frontal, frontotemporal, and basal ganglion regions.

FDG PET measures of regional metabolic activity in a group of unipolar and bipolar patients indicate that ictal glucose use increases globally from baseline levels and then drops sharply immediately postictally, as reported in a study by Ackermann.24

Guze25 studied four patients prior to a course of ECT and again either one day after the last ECT treatment (N = 3) or 112 days after completing a course of ECT (N=I). There was no reduction of local metabolic rate for glucose in those patients studied one day after the completion of ECT In the one case studied 112 days after ECT and in a patient maintained on Tegretol, there was an increase in ECT glucose metabolism. Guze suggested that the lack of decrease in glucose metabolism in those patients studied one day after their last treatment represented a "floor" effect, because in this group's experience, the mid-frontal gyrus is hypometabolic in depressed patients. On the basis of the one patient studied 112 days after completing a course of ECT, the authors speculate that ECT leads to normalization of mid-frontal gyrus glucose metabolism.

To date, there have been few 133Xe, SPECT, and PET studies of changes in CNS function after administration of ECT The studies completed so far shed little light on how ECT works. This is in large part because standardization of CBF and metabolism in patients with depression has been poor. Prospective studies using 133Xe, IMP and HMPAO SPECT, and FDG PET are difficult to do since radiation is involved, and repeated exposure to radiopharmaceuticals would be necessary.

It is likely that during seizure activity, RCBF and FDG metabolism are increased and this may become suppressed postictally, but the degree of suppression and the duration of the suppression is unclear. The long-term effects of ECT on these parameters is also unclear. If depression is associated with reduced CNS metabolism in various areas, including the prefrontal area, and if the ultimate effect of ECT is to increase these parameters as suggested by Guze,25 then this may indicate a corrective action of ECT. However, the underlying mechanism for the decrease in metabolic and CBF activities seen in depression remains to be elucidated.


New technologies are teaching us new lessons about ECT and confirming old findings. Through MRI and CT studies, we learn that there is no evidence of persistent structural changes associated with the administration of a course of ECT. Transient changes occur with ECT in T1 relaxation values, but these reverse within a relatively short period and are not cumulative. Many elderly patients who are referred for ECT have a variety of abnormalities, including cortical atrophy, white matter lesions, caudate lesions, and basal ganglion lesions. These have not been shown to worsen after ECT. All of these findings argue against the claim that ECT causes neuronal damage.

CNS scanning may also help predict which patients will respond, not respond, or have complications with ECT. However, more studies are required to confirm the preliminary reports described here. Unfortunately, 133Xe inhalation, SPECT, and PET studies have not provided enough information to correlate changes in CNS function with the efficacious or deleterious effects of ECT. A preliminary report suggests that depressed patients have decreased metabolic activity in the prefrontal areas of the brain, which increases after completion of a series of ECT. However, these results have not been confirmed with larger studies.


1. Friedberg J. Shock treatment, brain damage, and memory loss: a neurological perspective. Am J Psychiatry. 1977: 134:1010-1014.

2. Owens DGC, Johnstone EC, Crow TJ, Frith CD. Jagoe JR. Kreel L. Lateral ventricular size in schizophrenia: relationship to the disease process and its clinical manifestations. Psychol Med. 1985; 15:27-41.

3. Andreason NC. Swayze V. Flaum M, Alliger R, Cohen G. Ventricular abnormalities in affective disorder: clinical and demographic correlates. Am J Psychiatry. 1990; 147:893-899.

4. Coffey CE, Weiner RD, Djang WT. et al. Brain anatomic effects of electroconvulsive therapy: a prospective magnetic resonance imaging study. Arch Gen Psychiatry. 1991; 48:1013-1021.

5. Coffey CE, Fiegel GS, Djang WT, Cress M, Saunders WB, Weiner RD. Leukoencephalopathy in elderly depressed patients referred for ECT. Biological Psychiatry. 1988: 24:143-161.

6. Kolbiensson H, Arnaldsson OS, Petursson H, Skulason S. Computed tomographic scans in ECT patients. Acta Psychialr Scand. 1986;73:28-32.

7. Weinberger DR, Torrey EF, Neophytides AN, Wyatt RJ. Lateral cerebral ventricular enlargement in chronic schizophrenia. Arch Gen Psychiatry. 1979; 36:735-738.

8. Calloway SP, Dolan RJ, Jacoby RJ, Levy R. ECT and cerebral atrophy: a computed tomographic study. Acta Psychialr Scand. 1981; 64:442-445.

9. Nasrallah HA, McCalley-Whitters M, Pfohl V, Clinical significance of large ventricles in manic males. Psychiatry Res. 1984; 13:151-156.

10. Bergsholm P, Larsen JL, Rosendahl K, Holsten F. Electroconvulsive therapy and cerebral computed tomography: a prospective study. Acta Psychialr Scand. 1989; 80:566-572.

11. Mander AS, Whitfield A, Kean DM, Smith MA, Douglas RHB. Kendell RE. Cerebral and brain stem changes after ECT revealed by nuclear magnetic resonance imaging. ISI 1987; 69-71.

12. Scott AIF, Douglas RHB, Whitfield A, Kendell RE. Time course of cerebral magnetic resonance changes after electroconvulsive therapy. Br J Psychiatry. 1990; 156:551-553.

13. Pande AC, Grunhaus LJ, Aisen AM, Haskett RF. A preliminary magnetic resonance imaging study of ECT treated depressed patients. Biological Psychiatry. 1990; 27:102-104.

14. San dye R, Pardeshi R. The relationship between ECT and unresponsiveness and calcification of the pineal gland in bipolar patients. Int J Neurosci. 1990;54:301-306.

15. Figiel GS, Krishnan RR, Doraiswamy M. Subcortical structural changes in ECT-induced delirium. J Psychiatry Neurol. 1990:3:172-176.

16. Silkverskiold P, Gustafson L, Risberg J, Rosen I. Acute and late effects of electroconvulsive therapy: clinical outcome, regional cerebral blood flow and electroencephalogram. Ann N Y Acad Sci. 1986; 462:236-248.

17. Kanaya T, Yonekawa M. Regional cerebral blood flow in depression. Jpn J Psychiatry Neurol. 1990; 44:571-576.

18. Hurwitz TA, Clark C, Murphy E, Klonoff H, Martin WRW, Pate BW. Regional cerebral glucose metabolism in major depressive disorder. Can J Psychiatry. 1990; 35:684-688.

19. Baxter LR, Schwartz JM, Phelps ME, et al. Reduction of prefrontal cortex glucose metabolism common to three types of depression. Arch Gen Psychiatry. 1989; 46:243-250.

20. Raichle M, Taylor J, Herscovitch P, et al. Brain circulation and metabolism in depression. In: Greitz T, Ingvar DH, Widen L, eds. The Metabolism of the Human Brain Studies with Positron Emission Tomography. New York: Raven Press; 1985.

21. Baxter LR, Felps ME, Mazziotta JE, et al. Cerebral metabolic rates for glucose and mood disorders. Arch Gen Psychiatry. 1985; 42:441-447.

22. Silkverskiold P, Rosen I, Risberg J Effects of electroconvulsive therapy on EEG and cerebral blood flow and depression. Eur Arch Psychiatry Neurol Sci. 1987; 236:202-208.

23. Baje M, Medved V, Basic M, Topuzovic N, Babic D, Ivancevic D. Acute effect of electroconvulsive therapy on brain perfusion assessed by Tc99Mhexamethylpropyleneamineoxim and single photon emission computed tomography. Acta Psychiatr Scand. 1989; 80:421-426.

24. Ackermann RF. Engel J, Baxter L. Positron emission tomography and autoradiographic studies of glucose utilization during electroconvulsive seizures in humans and rats. Ann N Y Acad Sei. 1986; 462:263-269.

25. Guze BH. Baxter LR, Schwartz JM, Szuha MP, Liston EH. Electroconvulsive therapy and brain glucose metabolism. Convulsive Therapy. 1991; 7:15-19.


Sign up to receive

Journal E-contents