Date of Award
Medical Doctor (MD)
Rationale: Generalized spike-wave discharges (SWD) are the electroencephalographic (EEG) hallmarks of generalized seizures and epilepsies. These epileptiform discharges, best known in absence seizures, typically last for 3-10 seconds and often confer highly variable cognitive and/or behavioral impairments. Using the Repetitive Tapping Task (RTT) and the more attention-demanding Continuous Performance Task (CPT), previous investigations showed that the amplitude and duration of SWD on EEG may predict performance on these relatively simple tasks. The effect of generalized SWD on more complex behaviors like driving has been investigated in very limited studies with either small sample sizes or by using computerized driving games that do not adequately simulate realistic driving. Evidence from these studies suggests that generalized SWD prolong reaction time and impair drivers’ abilities to maintain car positions. At the same time, these generalized SWD may persist as subclinical epileptiform discharges even in “seizure-free” patients who continue to drive, posing a significant challenge to driver licensing authorities, clinicians and patients. These subclinical epileptiform discharges may pose transient cognitive impairments that presumably can lead to catastrophic injury and fatal motor vehicle crashes. Consequently, there is a need to characterize the impact of generalized SWD on driving and identify objective, clinically available predictors of impaired driving behavior during generalized SWD. We have established a practicable paradigm to test driving behavior during subclinical generalized SWD and to potentially identify EEG features of generalized SWD that impair driving.
Methods: Subjects 15 years or older, diagnosed with generalized epilepsy and with generalized SWD on ambulatory EEG but no clinical seizures in the preceding month, drive for an hour in a ½-cab high fidelity driving simulator (miniSimTM) equipped for real-time video and high-density (HD) EEG recording. A virtual road obstacle is manually presented every 5 minutes for baseline testing, as well as during SWD episodes. Subjects are instructed to safely pull over when the obstacle is presented. The simulator records more than 200 variables throughout the drive. Selected variables such as reaction time (ms), vehicle speed (mph), brake force (lb), and steering wheel velocity (deg/sec) are compared between periods of no SWD (as baseline) and periods with SWD (as test).
Results: We succeeded in recording reaction time (i.e., the time between obstacle presentation and application of brakes), brake force, vehicle speed, and steering wheel velocity at baseline and during 6 SWD in 2 subjects; a third subject did not have any SWD during testing. The road obstacle was successfully presented manually in 50% of all SWD that occurred during testing with a mean delay of 1.69 seconds from onset of discharges, which lasted an average of 5.9 sec (SD= 3.2sec). In 3/6 tested SWD, the subjects failed to respond to the road obstacle, whereas such omission happened in only 4/59 (6.78%) baseline trials (difference of proportions= 43.22%, p-value = 0.0012). Further, SWD during which subjects failed to appropriately respond to the road obstacle lasted longer on EEG than SWD during which subjects appropriately responded to the road obstacle (8.38 sec vs. 3.51 sec, p-value = 0.0402). The 6 SWD that were missed during testing were similar in duration to the 6 SWD that were successfully detected and tested (5.50 sec vs. 5.94 sec, p-value= 0.38336).
Conclusion: This study demonstrates that high-fidelity driving simulators instrumented for HD-EEG and video monitoring are feasible tools for studying driving behavior in people with epilepsy. This approach provides a previously unexplored avenue for studying driving safety among people with epilepsy, particularly in potentially identifying EEG features that may predict driving impairment in subclinical generalized SWD. Although limited, the results of the study demonstrate that subclinical SWD that impair driving last longer than those that spare driving, similar to what has been shown in other behavior tests (including the RTT and CPT). With further EEG and behavioral data collection, we anticipate the possibility of using machine-learning algorithms in classifying SWD as “sparing” vs. “impairing” driving ability.
Antwi, Prince, "High Fidelity Simulated Driving Paradigm For Predicting Driving Impairment In Generalized Epilepsy" (2019). Yale Medicine Thesis Digital Library. 3474.