Authors :
Presenting Author: Aziza Ysmanova, MD – Yale
Aya Khalaf, PhD – Yale
Matthew Gruen, BS – Yale University
Alexandre Castro, MD – Yale
Alberto Prat, MD – Yale
Basil Abdalla, MD – Yale
Nasrin Shahana, MD – Yale
Harry Sutherland, MD – Yale
Jay Liu, MD – Yale
Yang Zheng, MD, PhD – Yale
Lawrence Hirsch, MD – Yale University School of Medicine
Hal Blumenfeld, MD, PhD – Yale University
Rationale:
Accurate seizure detection is crucial for epilepsy diagnosis, treatment planning, and therapeutic evaluation. Patient awareness and the ability to self-report seizures vary significantly, influenced by seizure characteristics. This study investigates these associations using patient-initiated button press data recorded during video-EEG monitoring.
Methods:
We retrospectively reviewed 232 seizures from 56 adults (age ≥18) with medically refractory epilepsy who underwent long-term scalp video-EEG at a tertiary epilepsy center. Seizures were included if they had identifiable electrographic or behavioral onset and offset and were evaluated behaviorally. Seizure onset type, localization, lateralization, and EEG pattern were classified. Button presses were categorized as preictal, ictal, or postictal. Associations were evaluated using chi-square analyses.
Results:
Among patients, 31/56 (55.4%) had at least one button-press seizure. Of 232 seizures, 109 (47.0%) were associated with a patient-initiated button press. Of the button-press seizures, 71/109 (65.1%) occurred ictally, 32/109 (29.4%) postictally, and 6/109 (5.5%) preictally. Button presses occurred in 79/177 (44.6%) focal onset seizures, 4/17 (23.5%) generalized onset seizures, and 26/38 (68.4%) unknown-onset seizures (chi-squared = 5.92, p = 0.052). Presses were recorded in 57/126 (45.2%) temporal, 26/69 (37.7%) extra-temporal, and 26/37 (70.3%) seizures with unknown localization (chi-squared = 5.62, p = 0.060). By lateralization, button presses occurred in 50/115 (43.5%) right-sided, 19/47 (40.4%) left-sided, 9/15 (60.0%) bilateral, and 31/55 (56.4%) undefined-lateralization seizures (chi-squared = 2.30, p = 0.512). By EEG patterns, button presses were seen in seizures featuring rhythmic theta (36/88; 40.9%), delta (20/44; 45.5%), alpha (7/20; 35.0%), and beta activity (12/18; 66.7%), lateralized periodic discharges (8/12; 66.7%), focal spike-wave discharges (1/5; 20.0%), generalized spike-wave discharges (2/3; 66.7%), electrodecremental responses (0/11; 0.0%), and no EEG changes (23/31; 74.2%) (chi-squared = 14.87, p = 0.062). Overall, button presses were seen in 86/201 (42.8%) of seizures with EEG changes, and 23/31 (74.2%) seizures without EEG changes (chi-squared = 17.61, p = 0.00027).Conclusions:
The presence of EEG changes during seizures was strongly associated with a decreased probability of button presses by patients, suggesting that physiologically severe seizures more commonly interfere with the ability to self-report. No significant differences in button pressing were found based on seizure type, anatomical localization, or EEG pattern. Further research is needed to investigate if physiologically severe seizures, including those with impaired consciousness, are more likely to prevent self-report through button pressing. This work sheds light on factors determining patient self-awareness of seizures, guiding clinical care.
Funding:
K99NS133494