Gerla, Václav

AbstractObjectiveMost seizure forecasting algorithms have relied on features specific to electroencephalographic recordings. Environmental and physiological factors, such as weather and sleep, have long been suspected to affect brain activity and seizure occurrence but have not been fully explored as prior information for seizure forecasts in a patient‐specific analysis. The study aimed to quantify whether sleep, weather, and temporal factors (time of day, day of week, and lunar phase) can provide predictive prior probabilities that may be used to improve seizure forecasts.MethodsThis study performed post hoc analysis on data from eight patients with a total of 12.2 years of continuous intracranial electroencephalographic recordings (average = 1.5 years, range = 1.0–2.1 years), originally collected in a prospective trial. Patients also had sleep scoring and location‐specific weather data. Histograms of future seizure likelihood were generated for each feature. The predictive utility of individual features was measured using a Bayesian approach to combine different features into an overall forecast of seizure likelihood. Performance of different feature combinations was compared using the area under the receiver operating curve. Performance evaluation was pseudoprospective.ResultsFor the eight patients studied, seizures could be predicted above chance accuracy using sleep (five patients), weather (two patients), and temporal features (six patients). Forecasts using combined features performed significantly better than chance in six patients. For four of these patients, combined forecasts outperformed any individual feature.SignificanceEnvironmental and physiological data, including sleep, weather, and temporal features, provide significant predictive information on upcoming seizures. Although forecasts did not perform as well as algorithms that use invasive intracranial electroencephalography, the results were significantly above chance. Complementary signal features derived from an individual's historic seizure records may provide useful prior information to augment traditional seizure detection or forecasting algorithms. Importantly, many predictive features used in this study can be measured noninvasively.

Cardiac output (CO) assessment as a basic hemodynamic parameter has been of interest in exercise physiology, cardiology, and anesthesiology. Noninvasive techniques available are technically challenging, and thus difficult to use outside of a clinical or laboratory setting. We propose a novel method of noninvasive CO assessment using a single, upper-arm cuff. The method uses the arterial pressure pulse wave signal acquired from the brachial artery during 20-s intervals of suprasystolic occlusion. This method was evaluated in a cohort of 12 healthy individuals (age, 27.7 ± 5.4 yr, 50% men) and compared with an established method for noninvasive CO assessment, the open-circuit acetylene method (OpCirc) at rest, and during low- to moderate-intensity exercise. CO increased from rest to exercise (rest, 7.4 ± 0.8 vs. 7.2 ± 0.8; low, 9.8 ± 1.8 vs. 9.9 ± 2.0; moderate, 14.1 ± 2.8 vs. 14.8 ± 3.2 l/min) as assessed by the cuff-occlusion and OpCirc techniques, respectively. The average error of experimental technique compared with OpCirc was −0.25 ± 1.02 l/min, Pearson’s correlation coefficient of 0.96 (rest + exercise), and 0.21 ± 0.42 l/min with Pearson’s correlation coefficient of 0.87 (rest only). Bland-Altman analysis demonstrated good agreement between methods (within 95% boundaries); the reproducibility coefficient (RPC) = 0.84 l/min with R2= 0.75 at rest and RPC = 2 l/min with R2= 0.92 at rest and during exercise, respectively. In comparison with an established method to quantify CO, the cuff-occlusion method provides similar measures at rest and with light to moderate exercise. Thus, we believe this method has the potential to be used as a new, noninvasive method for assessing CO during exercise.