Mitigating Effects of Packet Loss in Wireless Neural Data Transfer: Exploring the Influence of Low-Cost Recovery Methods on Spikes and HFOs in iEEG

The wireless transmission of neural data may pose the risk of packet loss (PL), potentially compromising signal quality or, in extreme cases, causing complete data loss. Addressing lost packets is essential to ensure data integrity and preserve vital neural patterns. This study investigates the effect of PL interference on epilepsy neuro biomarkers, focusing specifically on interictal epileptiform spikes and high frequency oscillations (HFOs), and the performance of the low computational cost interpolation methods. We observed that 95% of spikes and 81% of HFOs could be recovered with linear interpolation at 5% PL, while 97% and 86%, respectively, with spline interpolation. Linear interpolation has the potential to recover neural events and reduce the noise floor with modest packet loss levels of up to 5% at a lower computational cost. However, for a higher level of PL, utilizing more intricate methodologies such as spline interpolation becomes imperative.