Thalamo-cortical evoked potentials during stimulation of the dentato-rubro-thalamic tract demonstrate synaptic filtering

Abstract

Essential tremor DBS targeting the ventral intermediate nucleus (Vim) of the thalamus and its input, the dentato-rubro-thalamic tract (DRTt), has proven to be an effective treatment strategy. We examined thalamo-cortical evoked potentials (TCEPs) and cortical dynamics during stimulation of the DRTt. We recorded TCEPs in primary motor cortex during clinical and supra-clinical stimulation of the DRTt in ten essential tremor patients. Stimulation was varied over pulse amplitude (2-10 ​mA) and pulse width (30-250 ​μs) to allow for strength-duration testing. Testing at clinical levels (3 ​mA, 60 ​μs) for stimulation frequencies of 1-160 ​Hz was performed and phase amplitude coupling (PAC) of beta phase and gamma power was calculated. Primary motor cortex TCEPs displayed two responses: early and all-or-none (<20 ​ms) or delayed and charge-dependent (>50 ​ms). Strength-duration curve approximation indicates that the chronaxie of the neural elements related to the TCEPs is <200 ​μs. At the range of clinical stimulation (amplitude 2-5 ​mA, pulse width 30-60 ​μs), TCEPs were not noted over primary motor cortex. Decreased pathophysiological phase-amplitude coupling was seen above 70 ​Hz stimulation without changes in power spectra and below the threshold of TCEPs. Our findings demonstrate that DRTt stimulation within normal clinical bounds does not excite fibers directly connected with primary motor cortex but that supra-clinical stimulation can excite a direct axonal tract. Both clinical efficacy and phase-amplitude coupling were frequency-dependent, favoring a synaptic filtering model as a possible mechanism of action.

Keywords: Deep brain stimulation; Essential tremor; Evoked potential; Phase-amplitude coupling.

Fig. 1

Experimental design using DBS lead stimulation and subdural electrodes for recordings. DBS leads are inserted into the DRTt just inferior to the Vim nucleus of the thalamus. Each lead contains four contacts, which can be stimulated in pairs using balanced square wave pulses of varying width and amplitude, and at frequencies over 150 ​Hz. Subdural electrodes are then visualized on a pial surface reconstructed from the T1 MRI; the DBS lead's entry point is displayed as a red sphere. A cutaway model of the brain shows the position of the DBS lead within the brain and fibers leaving the M1 SDEs passing by the DBS lead.

Fig. 2

Thalamo-cortical evoked potentials (TCEPs) are dependent on the charge density of stimulation. Pulse widths and amplitude were varied from clinical to supra-clinical (30–250 ​μs and 1.25–10 ​mA, all stimulation at 1 ​Hz) at a variety of amplitude-width pairs while controlling for charge density (2.5, 5, 10, and 20 ​μC/cm²). These parameters allowed for direct evaluation of the contribution of amplitude, pulse width, and charge density to TCEP morphology. Peaks were quantified by amplitude or prominence (or area-under-the-curve/AUC). The P1 and N2 TCEPs (first positive and second negative deflections) showed the greatest levels of variability. Scale bars for charge density graphs - N1: 20–60 ​μV and 50–100 ​μV∗ μs, P1: 30–80 ​μV and 60–135 ​μV∗ μs, and N2: 40–65 ​μV and 70–125 ​μV∗ μs.

Fig. 3

Phase amplitude coupling (PAC) decreases with increasing stimulation frequency. PAC was assessed as the relationship between beta (13–30 ​Hz) phase and gamma (30–200 ​Hz) amplitude. SDEs from six patients over M1 were included. PAC was high at baseline and during low frequency (<70 ​Hz) stimulation, but above 70 ​Hz resulted in a linear decrease in PAC. A baseline PAC from before DBS lead insertion is included for reference (dotted red line). The PAC for M1 SDEs was then compared with all non-M1 SDEs to estimate the significance (shown with ​± ​2 ​s.e.m., red dotted line denotes a t ​= ​0).

Fig. 4

Phase-amplitude coupling (PAC) by frequency. PAC plots (frequency by phase, frequency by amplitude) for different stimulation frequencies are shown for beta phase (10–30 ​Hz) and gamma amplitude (30–200 ​Hz). Results are averaged across the same electrodes used the PAC analysis (Fig. 4). There broadband decrease in PAC across beta-gamma coupling at stimulation above 70 ​Hz. PAC units are arbitrary.

Fig. 5

Power spectral density (PSD) estimates around stimulation. PSD was calculated using Welch's method (Hanning window of 256 ​ms, 50 ​% overlap) for the 2 ​s before stimulation (red, Pre-Stimulation), the 10 ​s during stimulation (black), and the 2s after stimulation stopped (blue, Post-Stimulation). PSDs for stimulation at 60, 70, 80, and 90 ​Hz are shown. Results are averaged across the electrodes used from the PAC analysis (Fig. 4). Each demonstrates a low frequency peak within the beta range (10–30 ​Hz), and a power-law decrease in the gamma range. Stimulation artifacts are seen at the frequency of stimulation and harmonics above that. Overall, the log (PSD) is stable in both the gamma and beta ranges.