Shared neurobiology
Upon analyzing dataset 1, the researchers found decreased gray matter in the bilateral anterior insula, dorsal anterior cingulate cortex, dorsomedial prefrontal cortex, thalamus, amygdala, hippocampus, and parietal operculum – findings that are “consistent with prior work.”
However, fewer than 35% of the studies contributed to any single cluster; and no cluster was specific to psychiatric versus neurodegenerative coordinates (drawn from dataset 2).
On the other hand, coordinate network mapping yielded “more statistically robust” (P < .001) results, which were found in 85% of the studies. “Psychiatric atrophy coordinates were functionally connected to the same network of brain regions,” the researchers reported.
This network was defined by two types of connectivity, positive and negative.
“The topography of this transdiagnostic network was independent of the statistical threshold and specific to psychiatric (vs. neurodegenerative) disorders, with the strongest peak occurring in the posterior parietal cortex (Brodmann Area 7) near the intraparietal sulcus,” the investigators wrote.
When lesions from dataset 3 were overlaid onto the ALE map and the transdiagnostic network in order to evaluate whether damage to either map correlated with number of post-lesion psychiatric diagnosis, results showed no evidence of a correlation between psychiatric comorbidity and damage on the ALE map (Pearson r, 0.02; P = .766).
However, when the same approach was applied to the transdiagnostic network, a statistically significant correlation was found between psychiatric comorbidity and lesion damage (Pearson r, –0.21; P = .01). A multiple regression model showed that the transdiagnostic, but not the ALE, network “independently predicted the number of post-lesion psychiatric diagnoses” (P = .003 vs. P = .1), the investigators reported.
All four neurosurgical ablative targets for psychiatric disorders found on analysis of dataset 4 “intersected” and aligned with the transdiagnostic network.
“The study does not immediately impact clinical practice, but it would be helpful for practicing clinicians to know that psychiatric disorders commonly co-occur and might share common neurobiology and a convergent brain network,” Dr. Taylor said.
“Future work based on our findings could potentially influence clinical trials and clinical practice, especially in the area of brain stimulation,” he added.
‘Exciting new targets’
In a comment, Desmond Oathes, PhD, associate director, Center for Neuromodulation and Stress, University of Pennsylvania, Philadelphia, said the “next step in the science is to combine individual brain imaging, aka, ‘individualized connectomes,’ with these promising group maps to determine something meaningful at the individual patient level.”
Dr. Desmond Oathes
Dr. Oathes, who is also a faculty clinician at the Center for the Treatment and Study of Anxiety and was not involved with the study, noted that an open question is whether the brain volume abnormalities/atrophy “can be changed with treatment and in what direction.”
A “strong take-home message from this paper is that brain volume measures from single coordinates are noisy as measures of psychiatric abnormality, whereas network effects seem to be especially sensitive for capturing these effects,” Dr. Oathes said.
The “abnormal networks across these disorders do not fit easily into well-known networks from healthy participants. However, they map well onto other databases relevant to psychiatric disorders and offer exciting new potential targets for prospective treatment studies,” he added.
The investigators received no specific funding for this work. Dr. Taylor reported no relevant financial relationships. Dr. Oathes reported no relevant financial relationships.
A version of this article first appeared on Medscape.com.