Deep brain stimulation creates informational lesion through membrane depolarization in mouse hippocampus

Abstract

Deep brain stimulation (DBS) is a promising neuromodulation therapy, but the neurophysiological mechanisms of DBS remain unclear. In awake mice, we performed high-speed membrane voltage fluorescence imaging of individual hippocampal CA1 neurons during DBS delivered at 40 Hz or 140 Hz, free of electrical interference. DBS powerfully depolarized somatic membrane potentials without suppressing spike rate, especially at 140 Hz. Further, DBS paced membrane voltage and spike timing at the stimulation frequency and reduced timed spiking output in response to hippocampal network theta-rhythmic (3-12 Hz) activity patterns. To determine whether DBS directly impacts cellular processing of inputs, we optogenetically evoked theta-rhythmic membrane depolarization at the soma. We found that DBS-evoked membrane depolarization was correlated with DBS-mediated suppression of neuronal responses to optogenetic inputs. These results demonstrate that DBS produces powerful membrane depolarization that interferes with the ability of individual neurons to respond to inputs, creating an informational lesion.

Fig. 1. Single-cell SomArchon fluorescence voltage imaging enables artifact-free neural recordings during DBS.

a Illustration of the experimental setup, the optical imaging window, and the animal preparation. The schematic was adapted from Piatkevich et al. 2019. b Schematic representation of a recorded CA1 neuron in the electric field generated by the electrode a few hundred microns away, and an example empirically measured electrical stimulation waveform. Lower right is an example imaging field showing GFP fluorescence from neurons expressing SomArchon-GFP and CoChR and the shadow of the nearby electrode. Scale bar 500 µm. The schematic neuron was adapted from Scidraw (doi.org/10.5281/zenodo.3925905). c Example SomArchon fluorescence before, during and after 40 Hz DBS. Average SomArchon fluorescence of an example neuron is shown in the upper left image. Scale bar, 15 µm. SomArchon trace is shown in black, detected spikes are marked by black ticks and electrical stimulation pulse patterns are in gold. SBR is the spike-to-baseline ratio, defined as the average spike amplitude divided by the standard deviation of the subthreshold SomArchon trace fluctuation for each recorded neuron (see methods). d Same as (c), but for 140 Hz DBS. e-f Same as (c-d), but other single neuron trial examples.

Fig. 2. DBS induces powerful membrane depolarization at both 40 Hz and 140 Hz.

a-c Amplitude of the subthreshold membrane potential (Vm) and corresponding quantifications. a, b Population-averaged Vm during 40 Hz DBS (a, n = 22 neurons) and 140 Hz DBS (b, n = 26 neurons). Vm is calculated as the fluorescence at each time point divided by the averaged amplitude of all spikes detected in a recording session for a given neuron. DBS-induced Vm changes were computed as normalized Vm by subtracting the mean of the Vm during the pre-stimulation baseline period. Shaded gray area is the standard error of the mean (SEM). c Quantification of Vm change from baseline during the transient (0–0.15 s) and sustained (0.15–1 s) periods of 40 Hz and 140 Hz DBS. Data are visualized as violin plots with the outer shape representing the data kernel density and box plots showing interquartile range (1x, 1.5x). The white lines in the boxes are the mean. For 40 Hz DBS, paired t-test, df = 21 neurons, transient: p = 0.002; sustained: p = 0.003, transient vs. sustained: p = 0.0156. For 140 Hz DBS, paired t-test, transient: p = 3.1 × 10⁻⁴, df = 25; sustained: p = 0.001, transient vs. sustained: p = 0.78. Comparison between 40 Hz and 140 Hz DBS, independent t-test, df = 46, transient: p = 0.045; sustained: p = 0.79. d-f Spike rate and corresponding quantifications. d, e Population-averaged spike rate during 40 Hz DBS (d, n = 22) and 140 Hz DBS (e, n = 26). Shaded gray area is SEM. f Violin plots of spike rate change from baseline during the transient and sustained periods of 40 Hz and 140 Hz DBS. For 40 Hz DBS, paired t-test, df = 21 neurons, transient: p = 0.0013; sustained: p = 0.002, transient vs. sustained: p = 0.19. For 140 Hz DBS, paired t-test, transient: p = 0.0071, df = 25; sustained: p = 0.11, transient vs. sustained: p = 0.01. Comparison between 40 Hz and 140 Hz DBS, independent t-test, df = 46, transient: p = 0.97; sustained: p = 0.089. g-i Time-frequency spectrum power of Vm and corresponding quantifications. Population-averaged Vm power during 40 Hz DBS (g, n = 22) and 140 Hz DBS (h, n = 26). i Violin plots of Vm 40 Hz or 140 Hz power change relative to the baseline during the transient and sustained periods of 40 Hz and 140 Hz DBS. For 40 Hz DBS, paired t-test, df = 21 neurons, transient: p = 1.05 × 10⁻⁴; sustained: p = 1.23 × 10⁻⁴, transient vs. sustained: p = 0.94. For 140 Hz DBS, paired t-test, transient: p = 9.77 × 10⁻⁶, df = 25; sustained: p = 0.0017, transient vs. sustained: p = 5.76 × 10⁻⁵. Comparison between 40 Hz and 140 Hz DBS, independent t-test, df = 46, transient: p = 4.87 × 10⁻⁴; sustained: p = 4.05 × 10⁻⁴. ns = non-significant, *<0.05, **<0.01, and ***<0.001. All paired or independent t-tests are two-sided. Source data are provided as a Source Data file.

Fig. 3. Individual electrical pulse evoked Vm and firing rate modulations are DBS frequency dependent.

a The population-averaged Vm trace before, during and after 40 Hz DBS with the pulse onset times indicated by the golden vertical lines. Green box indicates the time windows where zoom-in versions are shown in b, e. Shaded area represents SEM. Vm is calculated as the fluorescence at each time point divided by the average amplitude of all spikes detected in a recording session for a given neuron. DBS-induced Vm changes were computed as normalized Vm by subtracting the mean of the Vm during the prestimulation baseline period. b The population-averaged Vm trace (n = 22 neurons) around 40 Hz DBS onset. Shaded area represents SEM. Time window as indicated in (a). Time zero corresponds to the first DBS pulse time. Yellow dots represent Vm modulations during DBS that are two standard deviations from the baseline Vm distribution. c Same as (b), but showing the population-averaged firing rate. d–f Same as (a-c), but for 140 Hz DBS (n = 26 neurons). g Population-averaged Vm aligned to the onset of all pulses during 40 Hz stimulation. Golden lines indicate the pulse onset times. Shaded area represents SEM. h Same as (g), but for firing rate. i-j Same as (g, h), but for 140 Hz DBS.

Fig. 4. DBS entrains CA1 spiking at the stimulation frequency while reducing theta-rhythmic spike output.

a Example SomArchon trace (black) of a CA1 neuron with Vm highlighted blue (smoothed ± 6 ms), and spike times marked with red ticks. b Population phase-locking value (PLV) of spikes to Vm across frequencies. Spikes from all neurons were concatenated for PLV computation (40 Hz, n = 1972, 140 Hz, n = 1867). Black line represents spike-Vm PLV during baseline period and golden line represents spike-Vm PLV during 40 Hz DBS period. Shaded area is standard deviation estimated through bootstrapping. Frequency-axis is on a logarithmic scale. c Same as (b), but for 140 Hz DBS. d The population polar histogram of spike times relative to the phase of Vm filtered at 140 Hz using a Butterworth filter during the baseline period. e Same as (d), but during the 140 Hz DBS period. f Quantification of spike-Vm PLV at the DBS 40 Hz entrainment frequency during baseline (Base), 40 Hz DBS (Stim) and post-stimulation (Post) periods. Black error bars are standard deviation. Individual data points represent population mean samples obtained by bootstrapping. g Same as (f), but for 140 Hz DBS. h The circularly averaged population polar histogram of spike times relative to the phase of Vm filtered between 3–12 Hz (theta-frequency range). i Same (h), but during the 140 Hz DBS period. j Quantification of spike-Vm PLV in the theta-frequency range, averaged across 3–12 Hz, during baseline (Base), 40 Hz DBS (Stim) and poststimulation period (Post). Individual data points represent population mean samples obtained by bootstrapping. k Same as (j), but for 140 Hz DBS. Statistics are based on permutation testing comparing estimated values to permutation-derived null distributions, ns = non-significant, **<0.01, and ***<0.001. Source data are provided as a Source Data file.

Fig. 5. DBS evokes additional Vm depolarization without changing the spike rate in the presence of CoChR optogenetic inputs.

a Illustration of simultaneous CoChR-evoked membrane depolarization and SomArchon voltage imaging during DBS. Optogenetic blue light stimulation was delivered wide-field over an area of 350 microns defined by the 40x objective. Schematic neuron was adapted from scridraw.io (doi.org/10.5281/zenodo.3925905). b An example CA1 neuron’s SomArchon fluorescence trace (black) and spikes (black ticks) during 8 Hz CoChR activation (blue line) and 40 Hz DBS (gold line). 8 Hz CoChR activation occurred throughout the 3-second trial, whereas 40 Hz DBS occurred for 1 second in the middle of each trial. c Zoom-in view of the periods indicated by the dashed lines in (b), during the baseline (i) and the DBS period (ii). d Population-averaged Vm during simultaneous 8 Hz optogenetic activation (blue) and 40 Hz DBS (gold). Vm is calculated as the fluorescence at each time point divided by the average amplitude of all spikes detected in a recording session for a given neuron (n = 20). DBS-induced Vm changes were computed as normalized Vm by subtracting the mean of the Vm during the pre-stimulation baseline period. Shaded area represents SEM. e Same as (d), but with 140 Hz DBS (n = 21). f Quantification of the transient (0–0.15 sec) and the sustained (0.15–1 sec) Vm depolarization induced by either 40 Hz or 140 Hz DBS, in the presence of optogenetic activation. Data are visualized as violin plots with the outer shape representing the data kernel density and a box plot showing the interquartile range (1x, 1.5x). The white lines in the boxes are the mean. For 40 Hz DBS, paired t-test, df = 19, transient: p = 0.0036; sustained: p = 0.004, transient vs. sustained: p = 0.028. For 140 Hz DBS, paired t-test, df = 20; transient: p = 0.01, sustained: p = 0.0025, transient vs. sustained: p = 0.0187. Comparison between 40 Hz and 140 Hz DBS, independent t-test, df = 39, transient: p = 0.0082; sustained: p = 0.42. (g). Population-averaged firing rate during simultaneous 8 Hz optogenetic activation (blue) and 40 Hz DBS (brown, n = 16 neurons). Shaded area represents SEM. The purple line represents the further smoothed firing rate (300 ms rectangular smoothing). h Same as (g), but for population-averaged firing rate with 140 Hz DBS (n = 17). i Violin plots of the transient (0–0.15 sec) and the sustained (0.15–1 sec) firing rate changes relative to the baseline induced by either 40 Hz or 140 Hz DBS, in the presence of optogenetic activation. For 40 Hz DBS, paired t-test, df = 19, transient: p = 0.062; sustained: p = 0.28. For 140 Hz DBS, paired t-test, df = 20; transient: p = 0.76, sustained: p = 0.087. Given DBS conditions were not significant from baseline, we did no comparison across DBS conditions. Two-sided paired t-test for within DBS condition statistics and two-sided independent t-test for between DBS condition statistics. ns, non-significant, *<0.05, **<0.01, and ***<0.001. Source data are provided as a Source Data file.

Fig. 6. Individual electrical pulse evoked Vm and firing rate changes are differentially modulated by optogenetic depolarization.

a Population-averaged Vm aligned to individual electrical pulses during 40 Hz DBS with simultaneous optogenetic CoChR activation (blue line) and without CoChR activation (black line). Dashed lines indicate the onset of individual DBS electrical pulse times. Pulse-averaged Vm was first computed for each neuron and then averaged across neurons. Normalized Vm was then computed by subtracting the mean of the population pulse averaged Vm during the 10 ms (40 Hz DBS) or 5 ms (140 Hz DBS) time period before pulse onset. Shaded area represents SEM. b Electrical pulse evoked Vm change, computed as the difference between the maximum minus the minimum of the pulse averaged Vm per neuron. Data are visualized as violin plots with the outer shape representing the data kernel density and a box plot showing interquartile range (1x, 1.5x). The white lines in the boxes are the mean. Paired t-test, p = 3.3 × 10⁻⁵, df = 19. c, d Same as (a,b), but for population-averaged firing rate aligned to individual electrical pulses during 40 Hz DBS. Paired t-test, p = 0.7, df = 19 (e-h) Same as (a-d), but for 140 Hz DBS. Paired t-test (Vm), p = 0.088, df = 20. Paired t-test (Firing rate), p = 0.15, df = 20. Two-sided paired t-test for all tests. ns = non-significant and ***<0.001. Source data are provided as a Source Data file.

Fig. 7. DBS reduces a neuron’s ability to respond to somatic membrane depolarization inputs evoked optogenetically.

a Population-averaged Vm aligned to blue-light CoChR pulse onsets during baseline (back line) and 40 Hz DBS (golden line). Shaded area represents SEM. The vertical blue shaded area corresponds to the optogenetic CoChR activation period (42 ms). The average Vm during the baseline period (−50 ms to 0 ms) was subtracted from the Vm. b Quantification of 8 Hz CoChR-evoked Vm power change during baseline (Base), 40 Hz DBS (Stim) and post-stimulation period (Post). Data are visualized as violin plots with the outer shape representing the data kernel density and a box plot showing interquartile range (1x, 1.5x). The white lines in the boxes are the mean. Stim vs. Baseline, paired t-test, p = 5.22 × 10⁻⁷, df = 19. Stim vs. Post, paired t-test, p = 5.6 × 10⁻⁴, df = 19, Post vs. Baseline, paired t-test, p = 0.07, df = 19. c, d Same as (a-b), but for 140 Hz DBS condition. Stim vs. Baseline, paired t-test, p = 5.66 × 10⁻⁷, df = 20. Stim vs. Post, paired t-test, p = 0.0024, df = 20, Post vs. Baseline,, paired t-test, p = 6.22 × 10⁻⁴, df = 20. e Same as (a), but for firing rate. Firing rate was normalized by subtracting the average firing rate during the baseline period (−50 ms to 0 ms). f Violin plots of spike phase-locking value (PLV) to 8 Hz Vm during baseline (Base), 40 Hz DBS period (Stim) and post-stimulation period (Post). Only neurons included with >5spikes. Stim vs. Baseline, paired t-test, p = 0.008, df = 14. Stim vs. Post, paired t-test, p = 0.0065, df = 9, Post vs. Baseline, paired t-test, p = 0.49, df = 8. g, h Same as (e, f), but for 140 Hz DBS. Stim vs. Baseline, paired t-test, p = 0.0008, df = 11. Stim vs. Post, paired t-test, p = 0.0127, df = 9, Post vs. Baseline, paired t-test, p = 0.82, df = 10. Two-sided paired t-test for all tests. ns, non-significant, *<0.05, **<0.01, and ***<0.001. Source data are provided as a Source Data file.

Fig. 8. DBS-induced membrane depolarization predicts suppression of optogenetic inputs.

a A single neuron example showing the trial-averaged Vm during 8 Hz CoChR activation and 40 Hz DBS. The Vm has been smoothed with a 30 ms rectangular window to better highlight the depolarization effect in response to 8 Hz CoChR activation. This neuron exhibited prominent optogenetics-evoked Vm depolarization without DBS, and reduced Vm depolarization during 40 Hz DBS. Average SomArchon fluorescence of the example neuron is shown in the left image. Scale bar, 15 µm (b) Same as (a), but another neuron example with strong Vm depolarization during 40 Hz DBS. c The reduction of optogenetically induced 8 Hz Vm power is shown as a function of Vm change (DBS – Baseline) during 40 Hz (blue dots) and 140 Hz (red dots) DBS. Vm change is normalized by the average spike amplitude for each recorded neuron. Each dot represents a neuron (n = 41 neurons total). Neurons with stronger DBS-evoked Vm depolarization exhibit greater suppression of the 8 Hz optogenetically induced Vm power. The black line represents the fitted linear regression line. d The reduction of optogenetically induced 8 Hz spike-Vm phase locking value (PLV) is shown as a function of Vm depolarization amplitude during 40 Hz (blue dots) and 140 Hz (red dots) DBS (n = 28 total neurons). Neurons with less than 5 spikes during stimulation and baseline were excluded from this analysis. Source data are provided as a Source Data file.