EEG responses to TMS are sensitive to changes in the perturbation parameters and repeatable over time

Abstract

Background: High-density electroencephalography (hd-EEG) combined with transcranial magnetic stimulation (TMS) provides a direct and non-invasive measure of cortical excitability and connectivity in humans and may be employed to track over time pathological alterations, plastic changes and therapy-induced modifications in cortical circuits. However, the diagnostic/monitoring applications of this technique would be limited to the extent that TMS-evoked potentials are either stereotypical (non-sensitive) or random (non-repeatable) responses. Here, we used controlled changes in the stimulation parameters (site, intensity, and angle of stimulation) and repeated longitudinal measurements (same day and one week apart) to evaluate the sensitivity and repeatability of TMS/hd-EEG potentials.

Methodology/principal findings: In 10 volunteers, we performed 92 single-subject comparisons to evaluate the similarities/differences between pairs of TMS-evoked potentials recorded in the same/different stimulation conditions. For each pairwise comparison, we used non-parametric statistics to calculate a Divergence Index (DI), i.e., the percentage of samples that differed significantly, considering all scalp locations and the entire post-stimulus period. A receiver operating characteristic analysis showed that it was possible to find an optimal DI threshold of 1.67%, yielding 96.7% overall accuracy of TMS/hd-EEG in detecting whether a change in the perturbation parameters occurred or not.

Conclusions/significance: These results demonstrate that the EEG responses to TMS essentially reflect deterministic properties of the stimulated neuronal circuits as opposed to stereotypical responses or uncontrolled variability. To the extent that TMS-evoked potentials are sensitive to changes and repeatable over time, they may be employed to detect longitudinal changes in the state of cortical circuits.

Figure 1. Non-parametric statistical procedure to perform single-subject pairwise comparisons between TMS-evoked potentials.

Single-trial recordings from two different conditions (blue and red lines) were randomly mixed 1000 times (A) and averaged (B). Instantaneous distributions of averaged voltages were computed and centralized around zero by keeping record of the displacement δ(t) (C). The distribution of maximum absolute values of each centralized distribution was computed and used to define a significance threshold G as the (1-α)100^th percentile (D). Significance boundaries (gray dotted lines) were computed as (±G + δ(t)) and used to define the significantly different time samples (red stars) between conditions at a specific channel (E).

Figure 2. Results of pairwise comparisons between TMS-evoked potentials of a representative subject at the sensor (A,B) and at the source level (C,D).

Brain responses to stimulation of BA19 at I% intensity and 0° angle on day1 (blue traces) are compared with brain responses recorded during four different sessions (red traces), during which stimulation parameters were varied one at a time, namely stimulation site (BA6), intensity (I%+10%), angle (45°) and day (day8), resulting in 3 C comparisons and one NC comparison. For each comparison, superimposition of pairs of TMS-evoked potentials in all sensors is displayed in (A), while enlarged view of P1 channel is shown in (B), together with significance boundaries (dotted gray traces) and significantly different samples (red stars). Pairs of TMS-evoked cortical currents are shown in (C) as current density maps and in (D) as temporal profile of current density integrated over the left frontal and left occipital lobules, together with significantly different samples (red stars).

Figure 3. Divergence Index of all pairwise comparisons between TMS-evoked potentials.

Single DI values computed over the entire post-stimulus period (250 ms) are shown in (A) with the following color-coding: DIs of the C comparisons for changes in the stimulation site, intensity and angle are represented by cyan, black and green dots, respectively, while DIs of NC comparisons are depicted in yellow for same-day sessions and in red for one-week-apart sessions. DI values computed over different temporal windows of interest (0–60 ms, 60–120 ms, 120–250 ms) are reported in (B) with the same color-coding, except for NC comparisons that are summed together and plotted in orange.

Figure 4. ROC analysis applied to the DI. ROC curve is depicted as a solid black line, interspersed by blank dots representing the values of sensitivity and specificity (repeatability) associated to single DI values.

The optimal DI value of 1.67% computed according to the Younden index is shown as a black dot. Dashed line represents the ROC curve of a random classifier. Gray-shaded region indicates the area under the curve.