White Matter Integrity Predicts Recovery Response Time of Deep Brain Stimulation for Depression

17 

Abstract Submission 

Ha Neul Song¹, Helen Mayberg¹, Ki Sueng Choi¹

¹Icahn School of Medicine at Mount Sinai, New York, NY

Ha Neul Song  
Icahn School of Medicine at Mount Sinai
New York, NY

Helen Mayberg  
Icahn School of Medicine at Mount Sinai
New York, NY

Ki Sueng Choi  
Icahn School of Medicine at Mount Sinai
New York, NY

Deep brain stimulation (DBS) targeting the subcallosal cingulate cortex (SCC) has proven effective for treatment-resistant depression (TRD), across a set of consecutive cohorts. SCC is structurally connected with other brain regions via white matter (WM) bundles, and stimulation of all connections is crucial for DBS clinical outcomes. Despite consistent SCC targeting, which aims to maximize the activation of critical WM bundles, variability persists in the recovery response time across patients, potentially linked to baseline variations in brain abnormalities. This study explores the status of WM integrity (approximated using diffusion MRI) and their longitudinal changes in the critical WM activation pathways.

We assessed the time to reach a stable response (TSR, more than 50% improvement of HDRS17 in two consecutive weeks) in 33 TRD patients receiving SCC-DBS. Preoperative and longitudinal MRI data were acquired on a 3T scanner with 2 mm isotropic resolution and 60 directions, including five b0 and two opposite phase encoding directions for distortion correction. The study examined the relationship between TSR and baseline fractional anisotropy (FA) of targeted WM bundles (cingulum bundle, forceps minor, subcortical junction, and uncinate fasciculus). Pearson correlation between TSR and FA along each WM bundle's trajectory was conducted, selecting the area with the maximum r value. Selected FA values in WM bundles served as features for linear regression predicting TSR. In the same areas, longitudinal FA changes at 1, 3, and 6 months post-operation were analyzed to compare fast and slow responders (each n = 1).

Our findings reveal a significant negative correlation between TSR and FA in bilateral midcingulate cortex (left: r = -0.64, p < 0.01; right: r = -0.52, p < 0.01), bilateral forceps minor (left: r = -0.40, p = 0.02; right: r = -0.42, p = 0.02), and left uncinate fasciculus adjacent to left hippocampus (r = -0.46, p < 0.01) and left insula (r = -0.42, p = 0.01). A linear model of FA successfully predicted TSR (leave-one-out cross-validation; r= 0.68, r2 = 0.46, p < 0.001), with the left midcingulate cortex emerging as the strongest predictor among critical WM bundles. Moreover, post-hoc analysis found that the magnitude of FA increases in these regions over 6 months was associated with a faster response. Slow response was associated with FA values in cingulum bundles and forceps minor that did not change or decreased over the same period.

These findings suggest that WM abnormalities in critical WM bundles undergo repair with chronic SCC-DBS suggesting that DBS may facilitate neuroplasticity changes in selective activated WM pathways. This study sheds light on both sources of individual variability in SCC-DBS response time, as well as a potential mechanism mediating DBS antidepressant response.

Deep Brain Stimulation ¹

Psychiatric (eg. Depression, Anxiety, Schizophrenia) ²

White Matter Anatomy, Fiber Pathways and Connectivity

MRI

Psychiatric Disorders

Treatment

White Matter

Other - Deep Brain Stimulation; Depression