Intracranial Mapping of a Cortical Tinnitus System using Residual Inhibition

Abstract

Tinnitus can occur when damage to the peripheral auditory system leads to spontaneous brain activity that is interpreted as sound [1, 2]. Many abnormalities of brain activity are associated with tinnitus, but it is unclear how these relate to the phantom sound itself, as opposed to predisposing factors or secondary consequences [3]. Demonstrating "core" tinnitus correlates (processes that are both necessary and sufficient for tinnitus perception) requires high-precision recordings of neural activity combined with a behavioral paradigm in which the perception of tinnitus is manipulated and accurately reported by the subject. This has been previously impossible in animal and human research. Here we present extensive intracranial recordings from an awake, behaving tinnitus patient during short-term modifications in perceived tinnitus loudness after acoustic stimulation (residual inhibition) [4], permitting robust characterization of core tinnitus processes. As anticipated, we observed tinnitus-linked low-frequency (delta) oscillations [5-9], thought to be triggered by low-frequency bursting in the thalamus [10, 11]. Contrary to expectation, these delta changes extended far beyond circumscribed auditory cortical regions to encompass almost all of auditory cortex, plus large parts of temporal, parietal, sensorimotor, and limbic cortex. In discrete auditory, parahippocampal, and inferior parietal "hub" regions [12], these delta oscillations interacted with middle-frequency (alpha) and high-frequency (beta and gamma) activity, resulting in a coherent system of tightly coupled oscillations associated with high-level functions including memory and perception.

Copyright © 2015 The Authors. Published by Elsevier Ltd.. All rights reserved.

Figures

Figure 1

Summary of Experiment The experiment was performed on 2 days (upper and lower rows). On each day, the experiment was divided into two sessions (left and right columns), separated by a short break (gap between columns). Time is denoted by the horizontal axis. In each trial of the experiment, a masker was presented lasting 30 s (gray blocks), followed by four periods in which the subject rated the intensity of his tinnitus (green blocks) that were separated by three silent taskless recording periods of 10 s each (pale green/white blocks) whose data formed the basis of further analysis. Note that the duration of each session was not fixed, but rather depended on the sum of response times (green blocks) across a fixed number of repetitions. The rating scale used consisted of integers from −2 to +2 (though the range of responses given by the subject was −1 to 1, as shown here). A rating of 0 corresponded to the subject’s usual baseline tinnitus intensity, which he confirmed was the same as it had been immediately before the start of each experiment. Each recording period had a tinnitus rating value assigned to it that was the average of the ratings immediately preceding and following that period; these ratings were mainly −0.5 (average of −1, preceding, and 0, following) or 0. Data corresponding to the single rating of +1 (0 pre and +2 post) on day 2 were removed prior to analysis. As tinnitus ratings showed some correlation with overall time elapsed since the start of the experiment and with time elapsed since the end of the preceding masker, the tinnitus ratings were orthogonalized with respect to these variables, yielding partial tinnitus ratings. Tinnitus ratings corresponding to each recording period are shown in blue, and the corresponding partial ratings are shown in red. Note that partial tinnitus ratings have been centered to a mean of zero as part of the partialization process. See also Figure S1 for the subject’s audiological assessment.

Figure 2

Oscillatory Power and Phase Coherence Changes with Tinnitus Suppression Electrode locations are displayed with respect to a left lateral view of the convexity (A); coronal sections illustrating the location of depth electrode contacts with respect to the gray matter of Heschl’s gyrus (Ai; shown in dark gray); a superior view of the superior temporal plane centered on Heschl’s gyrus (Aii); the inferior surface of the temporal lobes (B); an axial slice through the anterior temporal lobes (Bi); an axial slice through the mid to posterior temporal lobe (Bii), with electrodes 173 and 174 in posterior parahippocampal cortex (PHC); and an axial slice through posterior temporo-occipital cortex, including the epileptogenic lesion (Biii). Axial slices in (Bi), (Bii), and (Biii) are viewed from their inferior aspects (in keeping with the view in B). In (Bi), the solid blue area indicates the lateral nucleus of the amygdala, and the solid yellow area the hippocampus proper, based on an automated segmentation algorithm implemented in FreeSurfer (https://surfer.nmr.mgh.harvard.edu/). In (Biii), the red area represents the lesion, and the gray area (also in A) represents the extent of tissue subsequently resected. Subdural electrodes are represented by solid black circles and depth electrodes by gray filled circles. Significant oscillatory power changes are denoted by colored hollow circles, with circle radius representing the peak correlation, within any of the specified frequency bands, between power and tinnitus suppression. Note that neural activity changes are displayed on all sections except (Ai), which is for illustrative purposes only. Blue, magenta, and orange circles indicate delta/theta/alpha (1–12 Hz) decreases, theta/alpha (4–12 Hz) increases, and beta2/gamma (28–148 Hz) increases, respectively. Phase coherence (PLV) changes were found only in the delta (1–4 Hz) band and are represented by solid blue lines. Each PLV connection was calculated between two bipolar electrode pairs (each consisting of two adjacent electrodes), and each end of each line is placed in between the two electrodes comprising that bipolar pair. Green boxes indicate sites where local cross-frequency coupling correlated significantly with tinnitus suppression (see Figure 4). Yellow and black boxes denote electrodes showing induced oscillatory and steady state responses to tinnitus-matched tones (see Figure S1B). See also Figure S2 for a full summary of power changes for both repetitions of the experiment.

Figure 3

Time-Frequency Decomposition of Oscillatory Power Changes Occurring during and after Noise Masker Presentation The horizontal axis represents time, from 5 s before masker onset, through the 30 s of masker presentation (gray box), and the 30 s following masker offset (during some of which the data constituting the main analyses were captured). Note that the next masker was presented much later than 30 s after the end of the present masker and that these plots simply present the time period during which most tinnitus suppression occurred. Vertical axes represent frequency in the same bands as featured in other analyses and figures. (A) and (B) show mean power changes, across trials, relative to pre-masker baseline (i.e., the 5 s before each masker), expressed as event-related spectral perturbations (ERSPs; 10 times the base-10 logarithm of the power to baseline ratio), grouped according to whether or not the first post-masker tinnitus rating indicated that RI was maintained at that time (RI trials) or had ceased (non-RI trials). Power values shown were averaged across all electrodes in auditory cortex (HG and STG). (C) shows T scores, for each time-frequency point, of the difference between RI and non-RI trials, thresholded for significance with a cluster approach at p 

Figure 4

Local Cross-frequency Coupling Changes Coinciding with Tinnitus Suppression Polar plots, in the complex plane, of significant changes in local cross-frequency envelope coupling as a function of tinnitus suppression. The five electrodes featured in this figure are highlighted with green boxes in Figure 2. In each plot (as summarized in E) the horizontal axis represents real-valued (R; non-lagged) coupling, with values to the right of the origin indicating positive coupling and to the left negative or anti-coupling. Real-valued envelope coupling is equivalent to envelope covariance, and thus an intuitive impression of the coupling results can be gained by just looking at the horizontal axis of each plot and ignoring the vertical. The vertical axis represents imaginary (I; phase-lagged) coupling, with values above the origin indicating the lower frequency leading and below the higher frequency leading. Distance from the origin indicates the strength of coupling. In each plot (as illustrated in F), the red dashed circle and line indicate the magnitude and phase difference, respectively, of coupling in the baseline tinnitus state. Blue dashed circles and lines indicate the magnitude and phase difference, respectively, of coupling in the suppressed tinnitus state. These are placed exactly on the equivalent red circles and lines in cases where magnitude or angle does not change significantly. These circles and lines indicate the state of coupling during the maximum partial suppression generally seen during the experiment. Bold arrows indicate the projected path of coupling with increasing tinnitus suppression, based on polar coordinate interpolation between the baseline and suppressed states. The number by each plot indicates the electrode number at which coupling is illustrated, and the Greek letters show which frequency bands the coupling being illustrated is between. HG, Heschl’s gyrus (A); PHC, parahippocampal cortex (B); STG, superior temporal gyrus (C); IPC, inferior parietal cortex (D). See also Figure S4, for a summary of cross-frequency coupling changes in the context of other neural activity changes.