Preserving Motor Features by Alternative Re-Referencing to Remove Heart Artifact on the Stentrode

Vascular electrocorticography (vECoG) has shown great promise as a less-invasive alternative to penetrating electrode arrays for neural signal acquisition. This study investigates the signal quality of the Stentrode device, a leading vECoG platform, by contrasting artifact persistence with the preservation of low frequency motor cortical activity. Typical (re-)referencing schemes, including a monopolar stent-mounted reference, a common average reference, and a Laplacian reference, are shown to significantly reduce the presence of electrocardiogram (ECG) artifacts compared to a distal monopolar reference. However, resting state beta activity is also significantly diminished when employing these techniques. By using Band-Limited Independent Component Analysis (BL-ICA), a type of spatial filter that allows weighting of the noise on each electrode differently, ECG artifacts are easily separated from vECoG recordings. With this cleaning methodology, signals are reconstructed without the ECG component and the reduction in cross-channel correlation is evaluated, as well as the increase in relative entropy between rest and go distributions. To ensure that low-frequency motor features are preserved, beta bursts features are evaluated in both the source space and reconstructed signal space. BL-ICA is an effective technique to remove widespread ECG artifacts while maintaining typical motor-related features in beta band activity.