Effects of stimulus pulse rate on somatosensory adaptation in the human cortex

Abstract

Background: Intracortical microstimulation (ICMS) of the somatosensory cortex can restore sensation to people with neurological diseases. However, many aspects of ICMS are poorly understood, including the effect of stimulation on percept intensity over time.

Objective: Here, we evaluate how tactile percepts evoked by ICMS in the somatosensory cortex of a human participant adapt over time.

Methods: We delivered continuous and intermittent ICMS to the somatosensory cortex and assessed the reported intensity of tactile percepts over time in a human participant. Experiments were conducted over approximately one year and linear mixed effects models were used to assess significance.

Results: Continuous stimulation at high frequencies led to rapid decreases in intensity, while low frequency stimulation maintained percept intensity for longer periods. Burst-modulated stimulation extended the time before the intensity began to decrease, but all protocols ultimately resulted in complete sensation loss within 1 min. Intermittent stimulation paradigms with several seconds between stimulus trains evoked intermittent percepts and also led to decreases in intensity on many electrodes, but never resulted in extinction of the sensation after over 3 min of stimulation. Longer breaks between each pulse train resulted in some recovery in the intensity of the stimulus-evoked percepts. For several electrodes, intermittent stimulation had almost no effect on the perceived intensity.

Conclusions: Intermittent ICMS paradigms were more effective at maintaining percepts. Given that transient neural activity dominates the response in somatosensory cortex during mechanical contact onsets and offsets, providing brief stimulation trains at these times may more closely represent natural cortical activity and have the additional benefit of prolonging the ability to evoke sensations over longer time periods.

Keywords: Adaptation; Brain-computer interfaces; Intracortical microstimulation; Microelectrode arrays; Sensory restoration; Somatosensory cortex.

Figure 1.. Continuous stimulation at higher frequencies resulted in faster adaptation.

A) Continuous frequency trains. Each line represents an individual pulse. B) The participant reported the perceived intensity of the ICMS-evoked sensations using an analog slider. The slider always started at a value of 1 and the participant moved the slider to indicate changes in perceived intensity. Each colored line represents the median intensity for 5 electrodes at a given frequency. The shaded regions represent the median absolute deviation. The vertical dotted lines indicate the end of stimulation for the 300 Hz train at 5 s and the end of stimulation for the 20 and 100 Hz trains at 15 s. Slider values after stimulation stopped are shown in a lighter shade to emphasize the effects during stimulation. C) Time at which the perceived intensity began to decrease for the mean response at each stimulus frequency. Error bars show the standard error and the dotted line indicates the maximum stimulation time of 15 s. * indicates p < 0.05 and ** indicates p < 0.001.

Figure 2.. Burst-modulated stimulation extinguished all percepts within 60 s.

A) Burst modulated trains. Each line represents an individual pulse. B) The participant indicated the perceived intensity of ICMS with an analog slider. The slider always started at a value of 1 and the participant moved the slider to indicate changes in perceived intensity. Each line represents the median intensity value across 10 tested electrodes for a given burst duration. The shaded regions represent the median absolute deviation. C-D) Bar plots showing C) the mean time at which the percept intensity began to decrease and D) the mean time at which the percept became undetectable. Error bars show the standard error. The dotted line indicates the maximum stimulation time of 60 s. * indicates p < 0.05.

Figure 3.. Intermittent stimulation caused less adaptation, which partially recovered over time.

A) Intermittent adaptation and recovery paradigms. The participant reported the perceived intensity on a self-selected scale immediately after each 1-s stimulus train. Each stimulus train was perceived as a distinct event. B) Mean normalized intensity reports over time. The perceived intensity values for each electrode were normalized to the value reported after the first train so that the data could be combined across all electrodes. Blue dots represent the median intensity value measured at each time point across all electrodes. Error bars show the standard error. (C) The raw (not normalized) intensity values reported for each electrode in two sessions during the adaptation and recovery periods. The participant used a self-selected scale so the reported values were not constrained to fall within a particular range. Data were not normalized so that differences in the initial intensity of different electrodes are highlighted. (D) The percent change in intensity for each electrode after the adaptation and recovery periods. The change in intensity is calculated from each electrode’s initial intensity. Circular markers indicate the first session while star markers indicate the second session.

Figure 4.. Stimulation frequency did not affect adaptation and recovery to intermittent stimulation on individual electrodes.

A) Using the intermittent stimulation protocol, four different electrodes were stimulated at 20, 100, and 250 Hz. Colored dots represent the mean intensity value measured at each time point across 4 test sessions. Error bars show the standard error. Different colors indicate the different stimulation frequencies. The dashed lines represent that each stimulus train was perceived as a discrete event. B) Intensity changes during adaptation and recovery at the three different frequencies. The raw change in intensity for the adaptation period was measured between the beginning and end of the adaptation period while the change for the recovery period was measured between the end of the adaptation period and end of the recovery period. Each bar shows the mean difference across four test sessions. Error bars show the standard error.

Figure 5.. Long periods of stimulation resulted in a slight increase in detection thresholds.

A) We measured the detection threshold and then applied stimulation for 15 s at 100 Hz followed by 15 s of no stimulation for four minutes and then measured the detection threshold. B) Detection thresholds before and after adaptation protocol. Each color represents a different electrode. C) The difference in detection thresholds before and after the adaptation paradigm. Colors are the same as in panel b. The dotted line represents the mean threshold difference.