A Case Study of Clinical and Neuroimaging Outcomes Following Repetitive Transcranial Magnetic Stimulation for Hoarding Disorder

To the Editor: We tested the effect of repetitive transcranial magnetic stimulation (rTMS) on hoarding symptoms and neuroimaging in a 58-year-old woman. The rTMS treatment included 30 sessions (5 days/week, 1 Hz, 90% resting motor threshold, 900 pulses/session) and targeted the right dorsolateral prefrontal cortex (rDLPFC, located with structural neuronavigation). Value-based decision-making deficits may underlie hoarding symptoms. In healthy volunteers, neuromodulation of the rDLPFC lowers item valuation (1) and alters a value-based decision-making network (along with the ventromedial prefrontal cortex [VMPFC] and amygdala) (e.g., 2). Thus, we hypothesized that rDLPFC neuromodulation would improve hoarding symptoms and alter activation and functional connectivity of the DLPFC, VMPFC, and amygdala.

A stable 3-week baseline was established (Saving Inventory Revised [SI-R] score range was 60–66). All rTMS sessions were completed on schedule with only mild and transitory side effects (e.g., headache). Symptoms improved after treatment (SI-R=46), and gains were maintained over 2 months (SI-R=45). On the Clinical Global Impressions Scale, the patient was rated as “minimally improved” at posttreatment. At follow-up (with no additional treatment initiated), she was rated as “much improved” and no longer met diagnostic criteria for hoarding disorder. The patient reported being “very satisfied” overall with the treatment.

The patient completed a computer simulation discarding task during functional magnetic resonance imaging at baseline, postbaseline, and posttreatment. The task entailed viewing pictures of household items presented on a screen. The patient indicated via finger press whether to keep or discard each item (for a detailed task description, see 3). Alternate versions were used to limit practice effects. The patient discarded more items over time (baseline=57%, postbaseline=67%, posttreatment=83%) and made discarding decisions faster (2,686 ms, 2,509 ms, and 1,955 ms, respectively). Compared with baseline, there was a decrease in VMPFC activation (p<0.001, k=10) but no activation changes in the DLPFC and amygdala. Functional connectivity between the rDLPFC stimulation point (a 5 mm diameter sphere) and the VMPFC (at activation, a 1 cm sphere) tended to increase (with a significant change in functional connectivity at posttreatment only), and the rDLPFC-left amygdala (mask defined by the FSL Harvard-Oxford atlas) functional connectivity tended to decrease. The VMPFC-left amygdala functional connectivity significantly increased at posttreatment (Figure 1).

This case demonstrates the potential for rTMS to treat hoarding disorder and suggests a neural mechanism of treatment. Results are consistent with research indicating that the VMPFC, and its connectivity with the DLPFC and amygdala, is critical to optimizing goal-directed choices during value-based decision-making and can be modulated using rTMS (e.g., 2). Replication of our results is needed using controlled designs, and generalization of treatment effects to other clinical populations characterized by decision-making deficits (e.g., obesity, addiction) should be explored.