Patients with severe treatment-resistant depression (TRD) who respond to subcallosal cingulate deep brain stimulation (SCC-DBS) can be distinguished from non-responders using machine learning models that incorporate pre- and post-surgical structural and metabolic brain measures.
Why this matters
TRD is a common and debilitating condition that results in high rates of morbidity and mortality. DBS is a treatment that has shown good success in several psychiatric conditions, and one in two patients with severe TRD will achieve long-term clinical improvement with SCC-DBS therapy.
Determining which patients with TRD will respond to SCC-DBS intervention is of critical importance. Investigators have examined several factors, including neurocognitive features, electroencephalogram activity, glucose metabolism, brain volume, and other neuroanatomical features, both at baseline and post-surgically.
This machine learning approach fed by a large, imaging-rich dataset of patients with SCC-DBS-treated TRD helps to clarify the features of responders and non-responders.