ECT is controversial. It is effective, but like any treatment we have in psychiatry, not everyone responds and there are side effects, including headache, transient confusion, and in some people short and long term problems with memory.

What clinicians would like to know is if there is a test which would predict who could respond to this and who would not. This would give those we advise with depression some certainty that this treatment would probably work.

A group used machine learning from MRI data to see if they could predict who could respond. The machine had a positive predictive rate of 78%: better than random. And one area in particular was associated with poor outcome: a larger volume of the cingulate gyrus.

Association of Pretreatment Subgenual Anterior Cingulate Gray Matter Volume (GMV) and Symptom Relief
A, Sagittal view depicts the positive association between the percentage of change in the Hamilton Depression Rating Scale (HDRS) score and subgenual anterior cingulate volume before electroconvulsive therapy (ECT) . Statistics were corrected for the entire brain volume , yielding only 1 significant cluster mapping to this area. B, Scatterplot depicts the positive correlation of the subgenual cingulate gyrus GMVs and the percentage of symptom improvement as measured by HDRS scores (ECT response). Contour lines indicate individual 95% CIs; central diagonal line indicates the association between subgenual anterior cingulate volume before ECT and percentage of change in the HDRS score.”

The authors note that every group that is working on this have varying findings and there is a need for replication. The equipment being used and the software being used is continuing to evolve. It is not, yet, reached the point where one should order an MRI and a machine prediction of outcome before ECT: the clinical utility of this approach is unknown.

Our analysis of the longitudinal effects of ECT confirmed previous findings that ECT induces massive structural plasticity, particularly in the hippocampus. Synaptogenesis, dendrogenesis, angiogenesis, or neurogenesis, as already shown in animal models, have been suggested as processes that mediate the observed effects.14 Furthermore, the results might reflect changes in blood flow or volume, although the anatomical reasons for local GMV changes remain unclear. However, in the present study, the extent of this hippocampal plasticity did not positively correlate to the extent of clinical response, as already reported, which indicates a possibility that these GMV increases are a byproduct rather than the underlying mechanism of ECT. On the other side, the missing link between GMV increases and symptom improvement may underlie long-term neurostructural effects that are associated not with acute but with latent symptomatic improvement. To date, whether structural changes can explain the therapeutic efficacy remains unclear, and the findings reported by recent studies are very inconsistent and often contradictory. The need for replication studies is urgent, particularly with regard to imaging measures and mood responses.

However, this, and other studies, show that the brain, under ECT, recovers. It stops looking different to people without depression. It becomes ‘non depressed’. The idea that ECT causes brain damage is obsolete. The newer imaging techiques, instead, suggest it causes brain recovery.