Responses of Cortical Neurons to Intracortical Microstimulation in Awake Primates

Abstract

Intracortical microstimulation (ICMS) is commonly used in many experimental and clinical paradigms; however, its effects on the activation of neurons are still not completely understood. To document the responses of cortical neurons in awake nonhuman primates to stimulation, we recorded single-unit activity while delivering single-pulse stimulation via Utah arrays implanted in primary motor cortex (M1) of three macaque monkeys. Stimuli between 5 and 50 μA delivered to single channels reliably evoked spikes in neurons recorded throughout the array with delays of up to 12 ms. ICMS pulses also induced a period of inhibition lasting up to 150 ms that typically followed the initial excitatory response. Higher current amplitudes led to a greater probability of evoking a spike and extended the duration of inhibition. The likelihood of evoking a spike in a neuron was dependent on the spontaneous firing rate as well as the delay between its most recent spike time and stimulus onset. Tonic repetitive stimulation between 2 and 20 Hz often modulated both the probability of evoking spikes and the duration of inhibition; high-frequency stimulation was more likely to change both responses. On a trial-by-trial basis, whether a stimulus evoked a spike did not affect the subsequent inhibitory response; however, their changes over time were often positively or negatively correlated. Our results document the complex dynamics of cortical neural responses to electrical stimulation that need to be considered when using ICMS for scientific and clinical applications.

Keywords: ICMS; neuronal excitability; nonhuman primate; short-term plasticity.

Figure 1.

Experimental setup and timeline. Macaques calmly sat in a chair receiving apple smoothie reward through the experiment. Cathodic, 200-μs phase width, single-pulse ICMS was delivered to one channel of the Utah array in primary motor cortex while unit responses were recorded across the array. Each session consisted of a prestimulus and stimulus epoch.

Figure 2.

Detection of evoked spikes and inhibition. A, Example of filtered data trace. The inset shows a stimulus followed by an evoked spike after 3.5 ms. B, Example PSTH (top) and corresponding raster plot (bottom). C, Defining evoked spikes. A PSTH with 0.5-ms bins was generated. Peaks after the time of stimulation greater than the upper threshold (mean + 2 SDs of −20 to −2 ms in the PSTH) down to the lower threshold (mean − 2 SDs) were called evoked spikes. D, Defining inhibition. Rather than using the PSTH, the inhibition duration was calculated for each stimulus by taking the time from stimulus onset to the next spontaneous spike. Stimulus intensities were 15 μA, unless stated otherwise.

Figure 3.

Stimulation during inhibition. Two representative examples of double-pulse and triple-pulse stimulation in which subsequent pulses arrive during the inhibitory response of the previous pulse (left). Spikes were readily evoked even when stimulating during the inhibitory response. Aligning the PSTHs to the final stimulus pulse (right) shows that the inhibition restarts at each stimulus pulse. Each condition consisted of 1500 stimuli. Bin width: 1 ms.

Figure 4.

Effect of distance from stimulus site. A, Evoked spike probability with respect to distance from the stimulated site in 0.25-mm bins. Stars (*) denote statistical significance from the closest group (ANOVA, p < 0.05). B, inhibition duration with respect to distance from the stimulated site in 0.25-mm bins. No groups were statistically different from the closest group (ANOVA). C, Probability histogram of evoked spike timings split into sites close (<1 mm, n = 121) to the stimulated site and all other sites (n = 299). The line shows the cubic interpolated moving average over three bins.

Figure 5.

Effect of stimulus amplitude. A, An example raster plot of a unit over different stimulus current amplitudes delivered for 5 min each at 2 Hz. Blue dots represent evoked spikes and the red line shows the median of the inhibition duration binned every 30 s with 1-s steps. B, An example evoked spike probability as a function of amplitude and (C) inhibition duration as a function of stimulus amplitude. The dashed lines are fitted sigmoidal curves using least-squares regression.

Figure 6.

Evoked spikes are not driven by an underlying state. A, Scatter plot showing the correlation coefficients of evoked spike activation for pairs of units with respect to distance between the recorded channels (n = 12,419 pairs). Red points show pairs of units that were significantly likely to be co-activated. Units that were farther away from each other were less likely to be co-activated. (Sig: statistically significant, p < 0.05; Not sig: not statistically significant). B, An example of distributions of stimuli that evoked a specific number of evoked spikes using the true data (Real) and when the evoked spikes were shuffled (Shuffled) for a pair of spikes. The two distributions are not statistically different (Kolmogorov–Smirnov test, p > 0.05). C, An example of distributions similar to B but extended to all evoked spikes within a single session. The two distributions are not statistically different (Kolmogorov–Smirnov test, p > 0.05).

Figure 7.

Probability of evoking a spike and inhibition duration are related to spontaneous firing rate. A, An example of a neuron with positively correlated firing rate (black) and evoked spike probability (blue) over 10 min (left), and a neuron with a negatively correlated firing rate (black) with inhibition duration (blue) over 10 min (right). The rate and probabilities are averaged over 30-s bins with 1-s steps. B, Scatter plot of the Pearson correlation coefficient (ρ) between the spontaneous firing rate and the probability of evoking spikes (left) or the inhibition duration (right) against the distance of the recorded unit from the stimulated site (n = 420). C, Distance from the stimulated site for units with uncorrelated, positively correlated, and negatively correlated evoked spike probabilities (left) or inhibition duration (right) with firing rate. Labeled p-values are from the Wilcoxon rank-sum test.

Figure 8.

Probability of evoking a spike depends on the timing of stimulus. Three examples of unit autocorrelations (Auto Cor), and probability of a stimulus evoking a spike relative to timing from the most recent spontaneous (Spont) and evoked (Evoked) spike. The plots show two different autocorrelation waveforms with correlated evoked spike probability. All traces show a moving average using 5-ms bins with a 1-ms step size. R_s is the correlation coefficient between Auto Cor and Spont, R_e the correlation coefficient between Auto Cor and Evoked, and p_s and p_e are the corresponding p-values of correlation.

Figure 9.

Changes in evoked spike probability and inhibition duration with repetitive stimulation. A, Left, Changes in evoked spikes across the array during a session with 10-Hz repetitive stimulation. A random unit was chosen for each electrode to demonstrate the lack of spatial organization of changes in responses. Right, Examples of changes in the probability of evoking spikes increasing or decreasing over time. B, Pearson correlation coefficients of evoked spike probability and inhibition duration plotted against stimulation frequency (n = 585). Note that experiments delivered tonic stimulation at 2, 5, 10, or 20 Hz; a jitter was added to the frequencies of each point to better visualize the data. C, The percentage of spikes that had significant (p < 0.05) changes over time for each stimulation frequency for both evoked spike probability and inhibition (left) as well as the ratio of decreases to increases (right). The dashed line of the right plot shows a threshold, if the value is higher (>1) the changes induced are more likely to be decreasing whereas if the value is lower (<1) the changes are more likely to be increasing. Numbers above points denote significance (1: significant from 2 Hz, 2: significant from 5 Hz, 3: significant from 10 Hz; p < 0.05, ANOVA).

Figure 10.

Changes with repetitive stimulation with respect to distance. A, Pearson correlation coefficients of evoked spike probability and inhibition duration plotted against distance of the recorded unit from the stimulated site (n = 585). Note that experiments delivered tonic stimulation at 2, 5, 10, or 20 Hz; a jitter was added to the frequencies of each point to better visualize the data. B, The percentage of spikes that had significant (p < 0.05) changes over time for each bin of distance (±0.25 mm around each point) for both evoked spike probability and inhibition (left) as well as the ratio of decreases to increases (right). We did not include data points from channels >2.75 mm from the stimulated site to this figure because of the lack of samples. Numbers above points denote significance (3: significant from 1.5 mm, 4: significant from 2 mm, 5: significant from 2.5 mm; p < 0.05, ANOVA).

Figure 11.

Changes in evoked spike latency with repetitive stimulation. A, An example of evoked spike latency changing over time. B, Scatter plot of Pearson correlation of the evoked spike latency over time with respect to stimulation frequency (left; n = 585). Percentage of evoked spike latencies with a significant change over time (middle), and the ratio of decreases to increases (right). Numbers above points denote significance (3: significant from 10 Hz, 4: significant from 20 Hz; p < 0.05, ANOVA). C, Scatter plot of Pearson correlation of the evoked spike latency over time with respect to distance from the stimulated site (left; n = 585). Percentage of evoked spike latencies with a significant change over time (middle), and the ratio of decreases to increases (right). Numbers above points denote significance (2: significant from 1 mm, 5: significant from 2.5 mm; p < 0.05, ANOVA).

Figure 12.

Relationship between evoked spikes and inhibition. A, Example PSTH with 1-ms bins following stimuli that evoked spikes and those that did not demonstrating similar inhibitory response. Each condition consisted of 1500 stimuli. Smoothed (2 ms wide Gaussian moving window) PSTH of the two different stimulation classifications. Note the inhibition is extremely similar for both. B, A comparison of the inhibition strength in the two different classifications for 470 units. There was no statistically significant pairwise difference between the two groups. C, Scatter plot of the Pearson correlation coefficient between evoked spike probability and inhibition duration against distance of the recorded unit from the stimulated site (n = 585). D, Comparisons of distance from the stimulus site of units with uncorrelated, positively correlated, and negatively correlated evoked spike probability and inhibition duration. The labeled p-value is from the Wilcoxon rank-sum test.

Figure 13.

Comparisons between regular spiking and fast spiking neurons. A, Example of regular spiking (RS) and fast spiking (FS) neuron waveforms. B, Distribution of the widths of spike waveforms (trough to peak time). Vertical dotted line indicates the classification boundary (0.35 ms) – fast spiking neurons fall to the left and regular spiking to the right. C, Spike width distribution of all recorded spikes and spikes that were evoked by stimuli. D, Distance from the stimulated site of evoked spikes grouped by spike width. E, Evoked spike probability distribution (left) and inhibition duration distribution (right) of RS and FS neurons. F, Spike width distributions of evoked spike probability change versus no change over time (left) and inhibition duration change versus no change over time (right). G, Spike width distributions of evoked spike probability decrease versus increase over time (left) and inhibition duration decrease versus increase over time (right).

Figure 14.

Stimulation response schematic. Schematic of cortical circuitry that generates excitatory and inhibitory ICMS responses through feedforward and feedback mechanisms. Stimulation activates axons projecting to the recording site. Sites closer to the stimulated site have more complete activation compared with sites further from the stimulated site. Possible direct connections from the stimulated site to the far site would also be sparsely activated.