Ongoing EEG artifact correction using blind source separation.

OBJECTIVE: Analysis of the electroencephalogram (EEG) for epileptic spike and seizure detection or brain-computer interfaces can be severely hampered by the presence of artifacts. The aim of this study is to describe and evaluate a fast automatic algorithm for ongoing correction of artifacts in continuous EEG recordings, which can be applied offline and online. METHODS: The automatic algorithm for ongoing correction of artifacts is based on fast blind source separation. It uses a sliding window technique with overlapping epochs and features in the spatial, temporal and frequency domain to detect and correct ocular, cardiac, muscle and powerline artifacts. RESULTS: The approach was validated in an independent evaluation study on publicly available continuous EEG data with 2035 marked artifacts. Validation confirmed that 88% of the artifacts could be removed successfully (ocular: 81%, cardiac: 84%, muscle: 98%, powerline: 100%). It outperformed state-of-the-art algorithms both in terms of artifact reduction rates and computation time. CONCLUSIONS: Fast ongoing artifact correction successfully removed a good proportion of artifacts, while preserving most of the EEG signals. SIGNIFICANCE: The presented algorithm may be useful for ongoing correction of artifacts, e.g., in online systems for epileptic spike and seizure detection or brain-computer interfaces.

The Onset of Interictal Spike-Related Ripples Facilitates Detection of the Epileptogenic Zone.

Objective: Accurate estimation of the epileptogenic zone (EZ) is essential for favorable outcomes in epilepsy surgery. Conventional ictal electrocorticography (ECoG) onset is generally used to detect the EZ but is insufficient in achieving seizure-free outcomes. By contrast, high-frequency oscillations (HFOs) could be useful markers of the EZ. Hence, we aimed to detect the EZ using interictal spikes and investigated whether the onset area of interictal spike-related HFOs was within the EZ. Methods: The EZ is considered to be included in the resection area among patients with seizure-free outcomes after surgery. Using a complex demodulation technique, we developed a method to determine the onset channels of interictal spike-related ripples (HFOs of 80-200 Hz) and investigated whether they are within the resection area. Results: We retrospectively examined 12 serial patients who achieved seizure-free status after focal resection surgery. Using the method that we developed, we determined the onset channels of interictal spike-related ripples and found that for all 12 patients, they were among the resection channels. The onset frequencies of ripples were in the range of 80-150 Hz. However, the ictal onset channels (evaluated based on ictal ECoG patterns) and ripple onset channels coincided in only 3 of 12 patients. Conclusions: Determining the onset area of interictal spike-related ripples could facilitate EZ estimation. This simple method that utilizes interictal ECoG may aid in preoperative evaluation and improve epilepsy surgery outcomes.

High frequency oscillations are less frequent but more specific to epileptogenicity during rapid eye movement sleep.

OBJECTIVE: We hypothesized that high frequency oscillations (HFOs) are differently suppressed during rapid eye movement sleep (REM) between epileptogenic and less epileptogenic cortices, and that the suppressive effect can serve as a specific marker of epileptogenicity. METHODS: Intracranial electroencephalography (EEG) was recorded in 13 patients with drug-resistant epilepsy. HFOs between 80 and 200Hz were semi-automatically detected from total 15-min EEG epochs each for REM and slow wave sleep (SWS). z-Score of HFO occurrence rate was calculated from the baseline rate derived from non-epileptogenic cortex. Intracranial electrodes were labeled as REM dominant HFO (RdH) if REM z-score was greater than SWS z-score or as SWS dominant HFO (SdH) if SWS z-score was greater than REM z-score. Relationship of electrode location to the area of surgical resection was compared between RdH and SdH electrodes. RESULTS: Out of 1070 electrodes, 101 were defined as RdH electrodes and 115 as SdH electrodes. RdH electrodes were associated with the area of resection in patients with postoperative seizure freedom (P<0.001), but not in patients without seizure freedom. CONCLUSIONS: HFOs near the epileptogenic zone are less suppressed during REM. SIGNIFICANCE: The less suppressive effect of REM may provide a specific marker of epileptogenicity.

Time-varying inter-hemispheric coherence during corpus callosotomy.

OBJECTIVE: Corpus callosotomy limits the bilateral synchrony of epileptic discharges. However, the instantaneous changes in bilateral synchrony during corpus callosotomy are unclear. The present study investigated how and when bilateral synchrony is suppressed in the anterior and then posterior steps of corpus callosotomy. METHODS: Intra-operative scalp electroencephalography (EEG) was recorded simultaneously with surgical video for six patients who underwent total corpus callosotomy for medically intractable drop attacks. The time-varying EEG inter-hemispheric coherence was quantified by wavelet transform coherence and trend analysis. RESULTS: The 4-13 Hz coherence decreased after corpus callosotomy in five patients. Significant decrease in coherence was observed only during the posterior step of callosal sectioning in three patients, but throughout both steps in two patients. CONCLUSIONS: Decrease in inter-hemispheric coherence is not always correlated with the stages of callosal sectioning. Inter-hemispheric coherence is decreased during the final stage of corpus callosotomy and the effect is maximized after sectioning is completed. SIGNIFICANCE: Various patterns of coherence decrease suggest individual variations in the participation of the corpus callosum in the genesis of bilateral synchrony. Time-varying inter-hemispheric EEG coherence is useful to monitor the physiological completeness of corpus callosotomy.