Ionic direct current modulation evokes spike-rate adaptation in the vestibular periphery

Abstract

Recent studies have shown that ionic direct current (iDC) can modulate the vestibular system in-vivo, with potential benefits over conventional pulsed stimulation. In this study, the effects of iDC stimulation on vestibular nerve fiber firing rate was investigated using loose-patch nerve fiber recordings in the acutely excised mouse crista ampullaris of the semicircular canals. Cathodic and anodic iDC steps instantaneously reduced and increased afferent spike rate, with the polarity of this effect dependent on the position of the stimulating electrode. A sustained constant anodic or cathodic current resulted in an adaptation to the stimulus and a return to spontaneous spike rate. Post-adaptation spike rate responses to iDC steps were similar to pre-adaptation controls. At high intensities spike rate response sensitivities were modified by the presence of an adaptation step. Benefits previously observed in behavioral responses to iDC steps delivered after sustained current may be due to post-adaptation changes in afferent sensitivity. These results contribute to an understanding of peripheral spike rate relationships for iDC vestibular stimulation and validate an ex-vivo model for future investigation of cellular mechanisms. In conjunction with previous in-vivo studies, these data help to characterize iDC stimulation as a potential therapy to restore vestibular function after bilateral vestibulopathy.

Figure 1

Afferent responses in the vestibular crista ex vivo are dependent on iDC stimulator position. (A) Schematic drawing of an anterior crista ampullaris with approximate location of iDC stimulator electrode and patch pipette. Modified from Tavazzani et al.. (B) Schematic drawing of the afferent innervation of type I and type II hair cells in the vestibular crista showing the position of the recording and stimulation electrodes in two configurations (inside and outside of the tissue). (C,D) Stimulus traces and representative afferent loose-patch recordings. When the iDC stimulator is outside the epithelium (C), there is an excitatory response to anodic (+) iDC steps and inhibitory response to cathodic (−) iDC steps (n = 9). When the iDC stimulator is inside the epithelium (D), the effect is reversed (n = 28).

Figure 2

Afferent responses are dependent on iDC stimulation amplitude. (A) CV and spontaneous rate distribution of the analyzed population. There was an inverse correlation between CV and spontaneous rate (Pearson’s; r = −0.7850; p < 0.0001; n = 21 recordings). (B) Individual (grey) and population (black) spike rates in response to anodic (+) and cathodic (−) 2 s-long iDC steps. iDC stimulation resulted in a change of the mean spike rate (Friedman; p < 0.0001; n = 21 recordings). Compared to the spontaneous spike rate (at 0 µA stimulus intensity), the anodic current steps reduced the spike rate, whereas cathodic current increased the spike rate. (C) CV and maximal spike rate change distribution of the analyzed recordings. Increased CV correlated to an increased maximal spike rate change (Pearson’s; r = 0.4879; p = 0.0248; n = 21 recordings) but this was not significant if the high CV outlier (circled) was excluded (p = 0.3182; n = 20 recordings). (D) Individual (grey) and population (black) spike rate change in response to anodic (+) and cathodic (−) iDC steps. Change in spike rate also varied significantly with change in amplitude (Friedman; p < 0.0001; n = 21 recordings). Cathodic iDC steps elicited a greater absolute spike rate change on average than anodic iDC steps (Wilcoxon; p < 0.0001; n = 21 recordings). Error bars are ± SEM. Dotted line at CV = 0.1 in (A,C) shows the cutoff between regular and irregular cells.

Figure 3

Characterization of firing rate and phase responses to sinusoidal iDC stimuli. (A) Stimulation waveform and representative afferent responses to a ±10 µA sinusoidal current across a range of frequencies. The shaded vertical column indicates the single cycle of the stimulus. The responses are aligned to the single stimulus cycle to observe the phase relationship of the responses to the different stimulation frequencies. (B) Enlarged 0.1 Hz responses from (A) with accompanying diary plot of instantaneous rate, to show the median spike calculation to determine excitatory (black notches)/inhibitory (grey notches) response phase relative to sinusoidal stimulus phase. (C–E) Individual (grey) and average (black) excitatory phase (C) and inhibitory phase (D) spike rate responses across stimulus frequencies. As frequency increased, the excitatory phase spike rate increased (Friedman; p < 0.0001; n = 15 recordings) and the inhibitory phase spike rate decreased (Friedman; p < 0.0001; n = 15 recordings). (E) Individual (grey) and average (black) phase shift of the excitatory spike rate response in relation to the cathodic stimulus waveform across different frequencies. There was a phase lag at low frequencies that decreased as stimulation frequency increased (Friedman; p > 0.0007; n = 15 recordings). Error bars are ± SEM. Representative traces were filtered with a Gaussian bandpass filter (5 Hz −2.5 kHz) to better display the data.

Figure 4

iDC stimulation induces spike rate adaptation. (A,B) Representative traces showing an afferent response to a +10 µA and −10 µA 60 s long iDC step, at 3 magnifications. The grey bar indicates the 50 ms period during the artifact where analysis was blanked. Dotted lines indicate representative spikes that were further expanded. (C,D) Diary plots of instantaneous spike rate versus time for the recordings shown in (A,B). Each data point reflects 1/inter-event interval. Red lines show one-phase exponential fits used to determine time constants of spike rate adaptation. (E) box and whiskers plot showing the spontaneous rate (60 seconds pre-stimulus), onset spike rate (first 50–550 ms after stimulus onset), and sustained spike rate (10–60 s after stimulus onset) during a 60 s long ±10 µA step. Compared to the baseline spike rate, at onset, anodic and cathodic steps reduced or increased the spike rate, respectively (Friedman; p < 0.0001; Dunn’s multiple comparisons; −10 µA onset: p = 0.28; +10 µA onset: p = 0.0061; n = 11 recordings) but sustained spike rates were not significantly different from baseline (Dunn’s multiple comparisons; −10 µA sustained: p > 0.9999; +10 µA sustained: p > 0.9999; n = 11 recordings). (F) Box and whiskers plots showing the distribution of spike rate adaptation time constants to 1 min ± 10 µA iDC steps. There was no significant difference in the time course of spike rate adaptation between cathodic and anodic steps (Wilcoxon; p = 0.5195; n = 11 recordings). In (A,B) for better display of the long duration data, the traces were filtered with a Bessel (8 poles) 50-Hz high pass filter.

Figure 5

Changes in iDC-induced spike rate activation after spike rate adaptation. (A,B) Individual (grey) and population (coloured) spike rate change in response to anodic (+) and cathodic (−) 2-s long iDC steps (protocol as in Fig. 2) after adaptation to a constant +10 µA (A) or −10 µA (B) baseline, applied for 60 s before steps. (C) Mean spike rate change in response to anodic (+) and cathodic (−) iDC across different baselines (+10 µA, 0 µA and −10 µA). There was a significant interaction between baseline and step amplitude (2way RM ANOVA; p = 0.0382; n = 10 recordings). At the greatest cathodic intensity (−20 µA), when compared to the 0 µA baseline control, the elicited change in spike rate was larger for the +10 µA baseline (Holm-Sidak multiple comparisons; p = 0.0016) and smaller for the −10 µA baseline (p = 0.0288). At the greatest anodic intensity (+20 µA), the change in spike rate was smaller for the +10 µA baseline (p = 0.0222) but not significantly different for the −10 µA baseline (p = 0.3129). Error bars are ±SEM.