A Computational Model of Deep-Brain Stimulation for Acquired Dystonia in Children

Abstract

The mechanism by which deep brain stimulation (DBS) improves dystonia is not understood, partly heterogeneity of the underlying disorders leads to differing effects of stimulation in different locations. Similarity between the effects of DBS and the effects of lesions has led to biophysical models of blockade or reduced transmission of involuntary activity in individual cells in the pathways responsible for dystonia. Here, we expand these theories by modeling the effect of DBS on populations of neurons. We emphasize the important observation that the DBS signal itself causes surprisingly few side effects and does not normally appear in the electromyographic signal. We hypothesize that, at the population level, massively synchronous rhythmic firing caused by DBS is only poorly transmitted through downstream populations. However, the high frequency of stimulation overwhelms incoming dystonic activity, thereby substituting an ineffectively transmitted exogenous signal for the endogenous abnormal signal. Changes in sensitivity can occur not only at the site of stimulation, but also at downstream sites due to synaptic and homeostatic plasticity mechanisms. The mechanism is predicted to depend strongly on the stimulation frequency. We provide preliminary data from simultaneous multichannel recordings in basal ganglia and thalamus in children with secondary dystonia. We also provide illustrative simulations of the effect of stimulation frequency on the transmission of the DBS pulses through sequential populations of neurons in the dystonia pathway. Our experimental results and model provide a new hypothesis and computational framework consistent with the clinical features of DBS in childhood acquired dystonia.

Keywords: basal ganglia; deep brain stimulation; dystonia; pediatric; single unit recording; thalamus.

Figure 1

Schematic model of the hypothesized propagation of signals at various frequencies of stimulation. (A) Predicted effect of stimulation below 2Hz. (B) Predicted effect of stimulation between 10 and 50Hz. (C) Predicted effect of stimulation between 100 and 150 Hz. (D) Predicted effect of stimulation above 200 Hz.

Figure 2

Smoothed spike rasters from right ventrolateral nucleus of the thalamus (VL) and two different 8-contact electrodes in right globus pallidus internus (GPi) during passive extension and flexion of the left knee in a child with static hemidystonia. Each raster shows eight different contacts from proximal (top) to distal (bottom) recorded over a 2 min period. Gray level is proportional to the total number of spikes in each 100 ms bin, normalized for each brain region: black represents the largest number of spikes seen in each region during 100 ms, while white corresponds to no spikes. Bottom trace shows EMG from left biceps (red), triceps (blue), flexor carpi radialis (aqua), and extensor carpi radialis (green) (Clinical data, previously unpublished).

Figure 3

Twenty-four hour extracellular microelectrode recording from a single neuron in Voa/Vop nucleus of the thalamus. Each row of the raster shows 1 h of recording. Hours from 17:00 until 16:00 the following day are indicated at the left. Gray level is proportional to the total number of spikes in each 20 s bin: black represents the largest number of spikes seen during 20 s, while white corresponds to no spikes (Clinical data, previously unpublished).

Figure 4

Stimulation at 10 Hz (top), 60 Hz (middle), and 185 Hz (bottom) in the simulated model. Each row represents a single simulated neuron [400 thalamus (blue), 100 inhibitory cortical (red), 400 excitatory cortical (green), and 400 motor output (black)]. Thalamus has sufficient intrinsic drive that it fires at high rates except for during the refractory period following a DBS pulse. The DBS frequency persists in cortex but is attenuated at the output.

Figure 5

Propagation of stimulus-related activity per stimulus pulse as a function of DBS stimulation frequency.

Figure 6

Propagation of non-stimulus-related background firing in thalamus as a function of DBS stimulation frequency.