In transcranial magnetic stimulation (TMS), a pulse of current driven through a coil on the scalp induces an electric field on the cortex, eliciting neuronal activation. Multi-locus TMS (mTMS) uses multiple coils to generate various electric field patterns, enabling electronic control of stimulation location and orientation. However, changing the stimulation target with mTMS can take up to a few seconds when the driving voltages are adjusted for each coil. To investigate whether multi-coil TMS combined with pulse-width modulation (PWM) in the microsecond scale could be used to overcome the issue of rapid pulse delivery to different cortical targets. We devised a methodology to generate and drive PWM approximations of given reference pulses. We compared resting motor thresholds (RMTs) and motor evoked potentials (MEPs) between trapezoidal and PWM pulses in several multi-coil targeting scenarios. The cortical electric field of PWM pulses had high temporal complexity, but the pulse types showed no statistically significant differences when stimulating with 2- or 3-coil combinations. With PWM pulses using five coils, the RMT increased by 9% (p=0.06), and MEP amplitudes decreased by 20% with stimulation doses over 120% RMT (p<0.05). Motor responses with trapezoidal and PWM pulses were equivalent when using two or three coils concurrently, but some differences were found with five coils. Using PWM with multi-coil TMS device to control the stimulation targets in millisecond-scale is feasible, although more research is needed to understand the neurophysiological effects of increasing coil count.

Published results of our Phase I safety and feasibility trial of accelerated intermittent theta burst stimulation (a-iTBS) in mild cognitive impairment (MCI) due to Alzheimer's disease showed a large effect-size improvement in cognition. Further demonstrate target engagement by identifying whether changes in local and network-level functional connectivity relate to the observed cognitive improvement. Eighteen patients with MCI received 3-day a-iTBS (8 sessions/day) to the left dorsolateral prefrontal cortex at Beam F3 (14,400 total pulses) and completed MRI and cognitive testing at pre- and post-treatment. Based on electric field models, we selected 3 stimulated target regions of interest (ROIs) which belonged to the frontoparietal (FPN), default mode (DMN), and ventral attention (VAT) networks (3 target networks). Metrics of resting-state functional connectivity were computed at the ROI level (within-network degree: number of connections) and network level (segregation: strength of connectivity within-network relative to other networks). We correlated changes in cognition and connectivity of the target ROIs and networks; off-target ROI (primary visual) and networks served as negative controls. Improvements in cognition were associated with connectivity changes in the target ROIs and networks, but not in off-target negative controls. Positive associations were observed for degree of the l-DMN and segregation of target networks overall, with significant effects for DMN and VAT. Cognitive improvement following a-iTBS in MCI may be attributable to local and network-level reconfigurations in functional connectivity. These findings will inform larger trials designed to further evaluate the neural mechanisms of a-iTBS for cognition in MCI.

Few treatments are available for individuals with marked treatment-resistant depression (TRD). Evaluate the safety and effectiveness of FDA-approved adjunctive vagus nerve stimulation (VNS) in patients with marked TRD. This 12-month, multicenter, double-blind, sham-controlled trial included 493 adults with marked treatment-resistant major depression who were randomized to active or no-stimulation sham VNS for 12 months. The primary outcome was percent time in response across months 3-12, with response defined as a ≥50% change from baseline on the Montgomery-Åsberg Depression Rating Scale (MADRS). Several secondary endpoints were evaluated. Overall, 88.4% of participants completed the trial. Percent time in MADRS response did not distinguish active from sham VNS. However, ratings from on-site clinicians (Clinical Global Inventory-Impression [CGI-I]), patients (Quick Inventory of Depressive Symptomology-Self Report [QIDS-SR]), and offsite masked raters (Quick Inventory of Depressive Symptomology-Clinician [QIDS-C]) revealed antidepressant benefits significantly favoring active VNS. Active VNS demonstrated significantly more percent time in response on the CGI-I (P=0.004) and QIDS-SR (P=0.049), and significantly more percent time in partial response (PR; symptom improvement ≥30%) on the CGI-I (P<0.001) and QIDS-C (P=0.006) versus sham VNS. Active VNS exceeded sham VNS in rate of dyspnea (P=0.035), a known side effect of VNS. No new adverse events were identified. Percent time in MADRS response did not distinguish the treatment groups, but on multiple instruments time in response and PR showed a positive treatment effect. VNS was found safe and effective in participants with marked TRD.

High-frequency repetitive transcranial magnetic stimulation (HF-rTMS) targeting the left dorsolateral prefrontal cortex (lDLPFC) is an established, clinically effective treatment for major depressive disorder (MDD). However, evidence of the cognitive effects of lDLPFC HF-rTMS, especially those cognitive functions affected by MDD, is mixed. We here assessed the cognitive effects of a single, offline, 10Hz rTMS session on task performance in an emotional faces N-back (EFNBACK) task, in both healthy and depressed individuals. We measured response times, accuracy, and sensitivity, in a sham-controlled, pre-post design. Importantly, using the Beck Depression Inventory (BDI-II), we assessed whether the cognitive effects of the stimulation are state-dependent on trait-state depression. We found lDLPFC HF-rTMS enhanced cognitive control over angry distractors. More importantly, these cognitive control effects were state-dependent on trait-state depression. HF-rTMS produced distinct performance changes dependent on baseline BDI-II scores. As a function of BDI-II scores, we observed either increased or decreased response times on the task. Further, we observed improved accuracy and sensitivity only on angry distractor trials as BDI-II scores increased, as a consequence of lDLPFC stimulation. These results underscore the role of HF-rTMS in enhancing executive control over negative emotional information by modulating lDLPFC, with effects varying according to depression state at the moment of the treatment session. HF-rTMS thus not only enhances inhibitory control over emotional stimuli but also exhibits such cognitive effects contingent on depressive state, contributing to our understanding of the state-dependence of therapeutic rTMS.

Motor responses evoked by transcranial magnetic stimulation (TMS) using posterior-anterior (PA) and anterior-posterior (AP) current directions have distinct latencies and thresholds. However, the underlying reasons for these differences remain unclear. To quantify the differences in activation sites between PA- and AP-TMS. Motor evoked potentials (MEPs) were recorded from five hand and arm muscles in nine healthy participants using both PA- and AP-TMS. Active motor thresholds were determined at 11 magnetic coil positions on the scalp. Probabilistic modelling was used to combine the measured threshold data with calculated electric field data from individual MRI-based models. This approach constructed 70 probability distributions of the activation site, dependent on the muscle and TMS direction. Modelling indicated that both PA- and AP-TMS more likely activated structures in white matter than in grey matter. PA-TMS activation sites were primarily in the white or grey matter in the precentral gyrus, while the AP-TMS activations were deeper and more posterior and lateral, likely within white matter under the postcentral and/or precentral gyri. Tractography and biophysical axon models provided a potential explanation on the location of activation sites: AP-TMS may activate the bends of white matter axons farther from M1 than PA-TMS, such that the conduction velocity along the neural tract could potentially explain the longer MEP latency of AP-TMS. The differences in activation sites among the five hand and arm muscles were small. While a direct experimental confirmation of the activation sites is still needed, the results suggest that electric field analysis combined with tractography and biophysical axon modelling could be a useful computational tool for analysing and optimizing TMS.