Skip Navigation LinksHome > June 2014 - Volume 30 - Issue 2 > Neuroimaging and Electroconvulsive Therapy: A Review
Journal of ECT:
doi: 10.1097/YCT.0000000000000140
Invited Reviews

Neuroimaging and Electroconvulsive Therapy: A Review

Bolwig, Tom G. MD, DMSc

Free Access
Article Outline
Collapse Box

Author Information

From the University of Copenhagen and Psychiatric Center Copenhagen, Denmark.

Received for publication January 13, 2014; accepted March 17, 2014.

Reprints: Tom G. Bolwig, MD, DMSc, Psychiatric Center Copenhagen, 2100 Copenhagen, Denmark (e-mail:

The author has no conflict of interest or financial disclosures to report.

Collapse Box


Background: Since the 1970s, a number of neuroimaging studies of electroconvulsive therapy (ECT) have been conducted to elucidate the working action of this highly efficacious treatment modality. The technologies used are single photon emission tomography, positron emission tomography, magnetic resonance imaging, magnetic resonance spectroscopy, and quantitative electroencephalography.

Methods: A PubMed literature search with focus on clinical studies was made from the inception of the database until December 2013 using the search terms electroconvulsive therapy and neuroimaging.

Results: Early methods allowing only identification of global changes of cerebral blood flow and cerebral metabolism show considerable ictal increases of these measures, which normalize during the postictal period. Later methodological developments have given access to measurements of minute activity changes in localized cortical and subcortical areas of the brain and have revealed differences in neurophysiology and metabolism between the hyperactive ictal state and the restorative interictal/postictal periods. Recent magnetic resonance imaging studies seem to pave way for new insights into ECT’s effects on increased connectivity in the brain during depression.

Conclusion: The existing data reveal considerable variations among studies and therefore do not yet allow the formulation of a unified hypothesis for the mechanism of ECT. The rapid developments in imaging technology, however, hold promises for further elucidation of the mode of action of ECT.

Since the introduction in 1938 of electroconvulsive therapy (ECT), this treatment modality has undergone a number of modifications; and the effects, beneficial and unwanted, have been the subject of numerous studies. However, we are not able to fully understand and analyze the events going on in the brain of patients receiving ECT. It is well documented that an induction of generalized seizures is a prerequisite for a beneficial therapeutic outcome. 1,2 This fact makes it highly difficult to point out singular events of relevance for the effects of ECT, as a generalized seizure elicits an earthquake of neurobiological events.

Understanding the therapeutic effects of ECT may well lead to an understanding of the pathophysiological causes of severe depression and other conditions responsive to ECT, and neuroimaging offers the opportunity to noninvasively measure changes throughout the brain both during and after ECT-induced seizures.

The present review will discuss some findings obtained by using a variety of neuroimaging methods including single photon emission computed tomography (SPECT), positron emission tomography (PET), and a variety of magnetic resonance imaging (MRI)-based techniques. These methods have helped throw light on especially cerebral blood flow (CBF) and cerebral metabolic rate (CMR) and, to a certain extent, on the presence of some transmitter substances and metabolites. In addition, some studies of modern electroencephalographic methods are discussed.

Mechanisms of ECT should be related to the pathophysiological causes of depression, and many neuroimaging studies point to both neurophysiological and structural abnormalities in patients with depression. Early xenon-133 inhalation studies thus found reduced global CBF as well as reduced regional CBF (rCBF) in selective frontal central superior temporal and anterior parietal regions in patients with severe depression compared with CBF in healthy controls. These findings seem to reflect a dysfunction in frontotemporoparietal cortical networks. 3 Later magnetic resonance spectroscopy (MRS) studies have shown increased glutamate levels, decreased occipital cortex γ-aminobutyric acid (GABA) concentrations, and changes in white matter of patients with severe depression compared to those of healthy controls. 4

Back to Top | Article Outline


Electroconvulsive therapy consists of ictal, postictal, and interictal periods. The boundaries between the postictal and interictal periods are vague and arbitrary because the postictal period is theoretically continuous with the interictal period. 5 It is well known that the biological changes occurring during generalized seizures are characterized by a general increase in neuronal electrical activity, neurotransmitter release, neuronal metabolism, and CBF in focal regions of the brain (rCBF), but also with some focal decreases. 6 Most studies point to a reversal of the changes observed in the ictal period, with general decreases in electrical activity, transmitter release, metabolism, and CBF and also with some focal increases.

Regarding the interictal period, this may display changes lasting for weeks to months before reversing.

Back to Top | Article Outline


Some early neuroimaging studies looked at rCBF using a xenon-133 clearance method and glucose and oxygen uptake using blood sample analyses during ECT-induced seizures. They showed an ictal doubling of rCBF, glucose metabolism, and oxygen consumption. 7,8 Later, Prohovnik et al used this technique and demonstrated that bilateral and unilateral electrode placement resulted in lateralized differences in rCBF.9

With the advent of SPECT, patients could be studied with the radioactive label tracer (99m)Tc-hexamethylpropylene amine oxime at some delay or immediately after a seizure begins. As the tracer does not redistribute to other areas of the brain, imaging can be done 60 to 90 minutes after seizure and a snapshot of CBF is captured at the time of the seizure. Using this technique, Blumenfeld et al 10 showed that selective focal networks in the brain involving frontal and parietal association cortices have increased CBF during generalized seizures regardless of bilateral or unilateral electrode placement. The same group looked at differences between different electrode placements (Blumenfeld et al 11 ) and found that bifrontal electrode placement resulted in increased rCBF in the prefrontal cortex and anterior cingulate gyrus while sparing the temporal lobes, whereas bitemporal electrode placement resulted in increased rCBF in lateral frontal cortex and anterior temporal cortex.

In contrast, unilateral electrode placement produced asymmetrical seizures with greater rCBF increases on the side of the stimulation, and this modification produced less disruption of verbal retrograde memory than was the consequence of bitemporal placement.

Elizagarate et al 12 also using (99m)Tc-hexamethylpropylene amine oxime SPECT found that CBF during the ictal period increased in cerebellum, brain stem, midbrain tegmentum, and other subcortical structures such as thalamus and basal ganglia; other groups measuring CBF and ECT found similar increases. 11,13 Enev et al 14 found that seizures initially elicit increase in rCBF in the regions of stimulation in the anterior frontotemporal cortex and also in subcortical structures such as the thalamus and basal ganglia. Likewise, Takano et al 15 using PET studies found that ictal rCBF increases both in subcortical structures such as the basal ganglia, brain stem, diencephalon, and amygdala and in the frontal, temporal, and parietal cortices compared with the rCBF before ECT.

In summary, rCBF, oxygen, and glucose uptake increase during seizures, the rise in rCBF being largest on the side of stimulation, yielding asymmetry with unilateral electrode placement. Some, but not all, studies find CBF increases in both cortical and different subcortical structures.

Back to Top | Article Outline


Most studies of the period after the seizure have been dealing with postictal rather than true interictal neuroimaging. Postictal imaging provides valuable information about the transition between acute treatment conditions and return to baseline. Imaging studies of this period have applied several methods from the xenon inhalation technique over SPECT, PET, MRS, and quantitative electroencephalography (EEG) (QEEG).

Most studies have found a general decrease in both rCBF and brain glucose metabolism in several brain regions immediately postictally. 9,16,17 Much like the findings with ictal studies, bilateral ECT was found to cause bilateral frontal hypoperfusion in the postictal period, whereas right unilateral ECT caused asymmetrical right greater than left postictal hypoperfusion. 16 Thus, regions that show the greatest hyperfusion ictally will show the greatest hypoperfusion postictally, in both cases reflecting activity changes in the anatomical regions where ECT seems to exert its immediately strongest effects.

Regional CBF studies using PET imaging showed increases in the thalamus and a decrease in the anterior cingulate and medial frontal cortex after the seizure period compared with pre-ECT measurements 14 and a decrease in glucose metabolism in both prefrontal regions. In addition, a decrease in rCBF after ECT was found in the left front polar gyrus, left superior temporal gyrus, left amygdala, nucleus accumbens, and globus pallidus in a SPECT study by Segawa et al. 17,18

The earliest xenon clearance studies found a 100% increase of rCBF during seizures that turned to baseline level 15 minutes later. 7 Also in the Prohovnik et al 9 study, rCBF normalized with a larger interictal decrease after bilateral electrode placement compared to unilateral placement. Silfverskiold et al 19 made similar observations, who also found that a decrease in EEG activity and that of rCBF do not always correlate. Acute and nonacute effects of EEG followed different patterns suggesting that CBF is more linked to cortical changes, whereas EEG is probably more related to activity in deeper subcortical structures.

A number of studies have looked at the CMRs for glucose after ECT. Most of these studies found a decrease in brain glucose metabolism in both frontal and parietal cortices and in the left temporal cortex 20,21 and a correlation between the decreased glucose metabolism and the clinical state as reflected in the Hamilton Depression Rating Scale scores were observed. 22 In a PET study, McCormick et al 23 found significant increases in metabolism in the left subgenual subregion of the anterior cingulate cortex and in hippocampus after ECT in responders; nonresponders had decreased values.

A literature review of studies by Schmidt et al 24 assessing possible changes in cerebral glucose metabolism before and after ECT by PET found many studies limited by small sample size, inhomogeneous study population, and methodological inconsistencies. Despite considerable variance, reduction in glucose metabolism after ECT in bilateral anterior and posterior frontal areas represented the most consistent findings. The same group looked for a potential correlation between clinical state and cerebral metabolic rates in patients receiving either bitemporal ECT or right unilateral ECT. They correlated the changes in relative metabolic rates in the left temporal lobe and left frontal basal regions, with changes in depression scores and neurocognition. Correlations were not significant. The finding of a decreased left frontobasal glucose metabolism is consistent with the limited data from previous studies, whereas the observation of increased left temporal glucose metabolism after ECT is a novel finding. The authors conclude that their study cannot support the view that glucose PET can assess the functional brain changes that are likely to occur subsequent to ECT. 24

Combining MRI and principles of nuclear magnetic resonance has led to the creation of MRS whereby spatial and temporal information about the brain and neurochemistry can be obtained.

Magnetic resonance spectroscopy allows observation of changes in metabolite turnover including glutamine, GABA, N-acetyl aspartate (NAA) (present predominantly in neuronal cell bodies where it acts as a neuronal marker), choline-containing compounds (Cho), creatine and phosphocreatine, glutamine/glutamate (Glx), lactate, and lipids.

Results from a number of studies show conflicting data. Thus, Sanacora et al 25 found that ECT induces an increase in cortical GABA concentrations, whereas Michael et al 26 found a significant increase of NAA only in patients who responded to ECT. In all successfully treated patients, parallel observations, meaning increased levels, were made for Glx, whereas Cho and creatine were unchanged. These findings suggest a neurotrophic effect of ECT. 26,27 For a review of ECT and neurotrophic effects, see Bouckaert et al 28 (this issue of JECT).

A study by Ende et al 29 showed no changes in the hippocampal NAA signals after ECT, but this group found an increase in the signal from Cho. The signals were observed bilaterally regardless of patients’ electrode placement. The authors conclude that ECT has not induced hippocampal atrophy or cell death. Choline-containing compound signals were significantly lower than normal before ECT and were normalized during ECT.

Obergriesser et al 30 made similar observations. They looked specifically at hippocampus and found no changes in hippocampal NAA signals after an interval of 20 months after the last ECT. The initially significant increase in the Cho signal was found to be reversed to nearly pre-ECT values. The increased Cho signals might reflect an increase membrane turnover and should reverse accordingly. The authors found that this increase in membrane turnover potentially might play a role in the therapeutic effect of ECT.

In summary, most studies of CBF and CMR for glucose find the increases elicited by the seizure reversed within minutes to hours. Right unilateral ECT was followed by greater hypoperfusion in the right frontal cortex, whereas bilateral ECT showed no such asymmetry. Despite conflicting MRS data findings of increased levels of Glx and Cho signals, both suggest a neurotrophic effect being behind ECT’s therapeutic effect.

Back to Top | Article Outline


Electroencephalography is a recording of brain wave activity; QEEG, popularly known as “brain mapping,” refers to a comprehensive analysis of brain wave frequency bandwidths that make up the raw EEG. Quantitative EEG provides complex analyses of a variety of brain wave characteristics. A well-extended EEG analysis is low-resolution electromagnetic tomography analysis (LORETA), which allows 3-dimensional (3D) viewing of electrical current sources in the brain. This methodology has a strong time resolution but a weaker power resolution than the other neuroimaging techniques and may be applied in real-time monitoring of brain activity.

Using LORETA data compared to an age-adjusted normative database, McCormick et al 31 studied patients with psychotic depression. They found that the theta band (4–7 Hz) activity was the only frequency band that changed with ECT. Analyses revealed the primary site of theta activity change being within the subgenual anterior cingulate cortex, a region showing also specific increase of blood flow and metabolism after ECT. 23 There was a positive association between increased theta wave activity and a decrease in psychotic symptoms, and the degree of low theta activity before ECT predicted the antipsychotic response to ECT.

Another LORETA study by Neuhaus et al 32 of depressed patients treated with ECT revealed a distinct pattern of activity in the depth of the temporal lobes during the interictal state of a bilateral ECT cycle. The authors suggested that the interictal focus of slow-wave activity may represent a correlate of episodic memory dysfunction during an ECT cycle, whereas they did not ascribe the EEG changes to be of any beneficial effect.

In summary, the few 3D electrophysiological imaging studies of ECT point to the potential of this technology to localize subcortical neuronal activity in real time and thus to demonstrate a correlation between neuronal activity and patients’ clinical condition.

Back to Top | Article Outline


Converging evidence from neuroimaging, neuropathological, and lesion analysis studies point to neural networks that modulate aspects of emotional behavior being implicated in the pathophysiology of mood disorders.

These networks involve the medial prefrontal cortex and closely related areas in the medial and caudolateral orbital cortex, amygdala, hippocampus, and parts of the basal ganglia, where alterations in gray matter volume and neurophysiological activity have been demonstrated in patients with recurrent episodes. 33 Further imaging studies of these areas in depressed patients point to a pathological “hyperconnectivity” among intracortical and corticolimbic circuits playing a role for the development of depression. Functional MRI (fMRI) studies thus found such abnormal network connectivity increased in depressed patients compared to healthy controls. 34,35

In a blood oxygenation level–dependent MRI study, Avery et al 36 compared depressed patients and control subjects during tasks requiring attention to visceral interoceptive sensations and examined between-group differences in insula resting-state functional connectivity. Severe depression was associated with greater resting-state functional connectivity between the dorsal midinsula cortex and limbic brain regions including the amygdala, subgenual prefrontal cortex, and orbitofrontal cortex. Connectivity between these regions and the dorsal midinsula cortex was positively correlated with depression severity.

Sheline et al, 37 also using fMRI, demonstrated that 3 different brain networks, the cognitive control network, default mode network, and affective network, all had increased connectivity to the same bilateral dorsal medial prefrontal cortex region, an area which the authors name the dorsal nexus. This nexus was found to have strongly increased depression-associated fMRI connectivity with large portions of each of the 3 networks. The authors suggested that reducing the increased connectivity might have an antidepressive effect.

The first study of ECT’s effect on increased activation to psychological tasks and decreased connectivity in frontal regions assumed to be related to emotional and cognitive efficiency during depression was conducted by Beall et al 38 in 6 depressed patients before and after treatment. Activation during both tasks was found clearly decreased after ECT, and remission of depression measured with the Hamilton Depression Rating Scale was significantly associated with reduced affective deactivation in the orbitofrontal cortex.

In an elegant fMRI study supporting the findings of Beall et al, Perrin et al 39 compared functional connectivity in 9 severely depressed patients before and after ECT. The treatment induced a significant decrease of functional connectivity in parallel with remission of depressive symptoms measured by the Montgomery-Åsberg Depression Rating Scale. The region with the largest decrease in functional connectivity was the left dorsolateral prefrontal cortical region. Both of these independent studies thus find ECT-induced reductions in connectivity in frontal regions in parallel with remission of the depression.

In summary, a number of fMRI studies point to an increased intrinsic connectivity in several neuronal networks implicated in different aspects of depression, perhaps representing both a biomarker for mood disorder and a potential therapeutic target. 39 Despite a limited number of patients studied, ECT’s clinical effects seem to correlate with a decreased connectivity, especially in frontal regions.

Back to Top | Article Outline


Neuroimaging has undoubtedly enhanced our knowledge of many physiological and neurochemical aspects of ECT. The methodology used in the 1970s and 1980s to measure in particular CBF and metabolism showed that seizures induced global and regionally different changes during and in the period after the seizure. New methods have given access to advanced measurements. Single photon emission computed tomography is a good method to perform ictal imaging because it can capture the rCBF at the time of the ECT-induced seizure, although it has lower spatial and temporal resolution compared with fMRI. Positron emission tomography is valuable for interictal imaging, as one can perform absolute measurements in vivo, in real time. Magnetic resonance spectroscopy provides data regarding biochemistry and function with very high spatial resolution.

In addition, modern EEG techniques may prove useful, as 3D images can be made using new algorithms. The power of resolution is weaker than found in PET and MRI, but 3D EEG’s temporal resolution is very strong and may be advantageously used in monitoring events taking place over longer periods of time.

Can we now, with the wealth of imaging data, formulate a unified hypothesis for the working action of ECT? The answer is no. There are still too many unanswered questions, and still, with the massive impulse traffic elicited throughout the brain, it is difficult to point out precisely what is relevant and what are epiphenomena.

Two areas of research applying modern imaging methods seem to have a strong potential to enhance our knowledge of the multitude of events taking place in the brain during and after seizures are induced. These are the study of neuronal impulse traffic with the help of 3D QEEG, which allows real-time registration of activity in all brain areas, and ECT’s effects on brain functional activation and connectivity.

Therefore, alongside other aspects of the rapid development of new technology within neuroscience advanced neuroimaging should give us even better tools than we have today for studying the mechanisms of ECT and most likely also throw more light on the pathophysiological causes of depression.

Back to Top | Article Outline


1. . Experimental studies of the mode of action of electroconvulsive therapy. Acta Psychiatr Scand . 1960;35:5–6.

2. . How does electroconvulsive therapy work? Theories in its mechanism. Can J Psychiatry . 2022;56:13–18.

3. , , , Regional cerebral blood flow in mood disorders. 1. Comparison of major depressives and normal controls at rest. Arch Gen Psychiatry . 1990;47:60–70.

4. , , , Increased cortical GABA concentrations in depressed patients receiving ECT. Am J Psychiatry . 2003;160:577–579.

5. , . Brain imaging and electroconvulsive therapy. In: . Electroconvulsive and Neuromodulation Therapies . New York: Cambridge University Press; 2009:94–108.

6. , . Epilepsy: A Comprehensive Textbook . 2nd ed. 2008:Philadelphia, PA: Lippincott Williams.

7. , , , Cerebral hyperemia in electrically induced epileptic seizures. Arch Neurol (Chicago) . 1973;28:335–342.

8. , , , The permeability of the blood–brain barrier during electrically induced seizures in man. Eur J Clin Invest . 1977;7:87–93.

9. , , , Acute reductions of regional cerebral blood flow following electroconvulsive therapy. Interactions with modality and time. Ann N Y Acad Sci . 1986;462:249–262.

10. , , , Selective frontal, parietal, and temporal networks in generalized seizures. Neuroimage . 2003a;19:1556–1566.

11. , , , Targeted prefrontal cortical activation with bifrontal ECT. Psychiatry Res . 2003a;123:165–170.

12. , , , Study of the influence of electroconvulsive therapy on the regional cerebral blood flow by HMPAO-SPECT. J Affect Disord . 2001;65:55–59.

13. , . Focal network involvement in generalized seizures: new insights from electroconvulsive therapy. Epilepsy Behav . 2004;5:3–12.

14. , , , Imaging onset and propagation of ECT-induced seizures. Epilepsia . 2007;48:238–244.

15. , , , (2007) Changes in regional cerebral blood flow during acute electroconvulsive therapy in patients with depression: positron emission tomographic study. Br J Psychiatry . 2007;190:63–68.

16. , , , Regional cerebral blood flow in mood disorders, III. Treatment and clinical response. Arch Gen Psychiatry . 1994;51:884–897.

17. , , , Regional cerebral blood flow changes in depression after electroconvulsive therapy. Psychiatry Res . 2006;147:135–143.

18. , , , Short-term effects of electroconvulsive treatment on the uptake of 99mTc-exametazime into brain in major depression shown with single photon emission tomography. J Affect Disord . 1994;30:27–34.

19. , , . Effects of electroconvulsive therapy on EEG and cerebral blood flow in depression. Eur Arch Psychiatry Neurol Sci . 1987;236:202–208.

20. , , , Decreased regional brain metabolism after ECT. Am J Psychiatry . 2001;158:305–308.

21. , , , Effects of electroconvulsive therapy on brain glucose metabolism: a preliminary study. Convuls Ther . 1988;4:199–205

22. , , , The effects of ECT on brain glucose: a pilot FDG PET study. J ECT . 2001;17:33–40.

23. , , , Metabolic correlates of antidepressant and antipsychotic response in patients with psychotic depression undergoing electroconvulsive therapy. J ECT . 2007;23:265–273.

24. , , , Changes in brain metabolism after ECT-positron emission tomography in the assessment of changes in glucose metabolism subsequent to electroconvulsive therapy—lessons, limitations and future applications. J Affect Disord . 2008;106:203–208.

25. , , , (2004) Subtype-specific alterations of gamma-aminobutyric acid and glutamate in patients with major depression. Arch Gen Psychiatry . 2004;61:705–713.

26. , , , Neurotrophic effects of electroconvulsive therapy: a proton magnetic resonance study of the left amygdalar region in patients with treatment-resistant depression. Neuropsychopharmacology . 2003;28:720–725.

27. , . Molecular and cellular actions on chronic electroconvulsive seizures. J ECT . 1998;14:181–193.

28. , , , ECT: its brain enabling effects. A review of electroconvulsive therapy–induced structural brain plasticity. J ECT . 2014;30:143–151.

29. , , , The hippocampus in patients treated with electroconvulsive therapy: a proton magnetic resonance spectroscopic imaging study. Arch Gen Psychiatry . 2000;57:937–943.

30. , , , Long-term follow-up of magnetic resonance-detectable choline signal changes in the hippocampus of patients treated with electroconvulsive therapy. J Clin Psychiatry . 2003;64:775–780.

31. , , , Antipsychotic effect of electroconvulsive therapy is related to normalization of subgenual cingulate theta activity in psychotic depression. J Psychiatr Res . 2009;43:553–560.

32. , , , Interictal slow-wave focus in left medial temporal lobe during bilateral electroconvulsive therapy. Neuropsychobiology . 2005;52:183–189.

33. , , . Brain structural and functional abnormalities in mood disorders: implications for neurocircuitry models of depression. Brain Struct Funct . 2008;213:93–118.

34. , , , Fronto-cingulate effective connectivity in major depression: a study with fMRI and dynamic causal modeling. Neuroimage . 2008;43:645–655.

35. , , , Aberrant functional connectivity of dorsolateral prefrontal and cingulate networks in patients with major depression during working memory processing. Psychol Med . 2009;39:977–987.

36. , , , Major depressive disorder is associated with abnormal interoceptive activity and functional connectivity in the insula. Biol Psychiatry . 2013. pii: S0006-3223(13)01070-6. doi: 10. 1016/j.biopsych.2013.11.027.

37. , , . Resting-state functional MRI in depression unmasks increased connectivity between networks via the dorsal nexus. Proc Natl Acad Sci U S A . 2010;107:11020–11025.

38. , , , Effects of electroconvulsive therapy on brain functional activation and connectivity in depression. J ECT . 2012;28:234–241.

39. , , , Electroconvulsive therapy reduces frontal cortical connectivity in severe depressive disorder. Proc Natl Acad Sci U S A . 2012;109:5464–5468.


electroconvulsive therapy; neuroimaging; hyperconnectivity

© 2014 by Lippincott Williams & Wilkins


Article Tools


Article Level Metrics

Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.