Integrating artificial intelligence with real-time intracranial EEG monitoring to automate interictal identification of seizure onset zones in focal epilepsy

Varatharajah, Yogatheesan; Berry, Brent M.; Cimbálník, Jan; Kremen, Václav; Gompel, Jamie Van; Stead, Matt; Brinkmann, Benjamin H.; Iyer, Ravishankar K.; Worrell, Gregory A.

doi:10.1088/1741-2552/aac960

Cited by 68 publications

(73 citation statements)

References 87 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…15,16 This promoted the development of combinations of several markers for achieving a better identification of the EZ. 17, 18 Cimbalnik et al proposed that a multifeature approach based on HFO rates, univariate and bivariate connectivity measures is superior to single marker approaches; surgical outcome was correctly predicted in 89% of patients when applying this approach to a 30-minute section of intracranial electroencephalography (iEEG). 19 Variations across the sleep-wake cycle are not negligible, when assessing markers of the EZ, 20-22 a fact often not taken into consideration.…”

Section: Introductionmentioning

confidence: 99%

NREM sleep is the state of vigilance that best identifies the epileptogenic zone in the interictal electroencephalogram

et al. 2019

Self Cite

View full text Add to dashboard Cite

Objective: Interictal epileptiform anomalies such as epileptiform discharges or highfrequency oscillations show marked variations across the sleep-wake cycle. This study investigates which state of vigilance is the best to localize the epileptogenic zone (EZ) in interictal intracranial electroencephalography (EEG). Methods: Thirty patients with drug-resistant epilepsy undergoing stereo-EEG (SEEG)/sleep recording and subsequent open surgery were included; 13 patients (43.3%) had good surgical outcome (Engel class I). Sleep was scored following standard criteria. Multiple features based on the interictal EEG (interictal epileptiform discharges, high-frequency oscillations, univariate and bivariate features) were used to train a support vector machine (SVM) model to classify SEEG contacts placed in the EZ. The performance of the algorithm was evaluated by the mean area under the receiver-operating characteristic (ROC) curves (AUCs) and positive predictive values (PPVs) across 10-minute sections of wake, non-rapid eye movement sleep (NREM) stages N2 and N3, REM sleep, and their combination.Results: Highest AUCs were achieved in NREM sleep stages N2 and N3 compared to wakefulness and REM (P < .01). There was no improvement when using a combination of all four states (P > .05); the best performing features in the combined state were selected from NREM sleep. There were differences between good (Engel I) and poor (Engel II-IV) outcomes in PPV (P < .05) and AUC (P < .01) across all states. The SVM multifeature approach outperformed spikes and high-frequency oscillations (P < .01) and resulted in results similar to those of the seizure-onset zone (SOZ; P > .05). Significance: Sleep improves the localization of the EZ with best identification obtained in NREM sleep stages N2 and N3. Results based on the multifeature classification in 10 minutes of NREM sleep were not different from the results achieved by the SOZ based on 12.7 days of seizure monitoring. This finding might ultimately result in a more time-efficient intracranial presurgical investigation of focal epilepsy. K E Y W O R D Sconnectivity, drug-resistant epilepsy, high-frequency oscillations, machine learning, sleep-wake cycle | 2405 KLIMES Et aL.

show abstract

Section: Introductionmentioning

confidence: 99%

NREM sleep is the state of vigilance that best identifies the epileptogenic zone in the interictal electroencephalogram

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…Even though the concept of the "EZ" has been used for more than 50 years (Bancaud et al, 1965), its precise definition remains controversial, and a reliable biomarker is still missing. Several algorithms have been proposed to assess the extent of the EZ and compare it with either the EZ area identified by an expert clinician or the resected region (Andrzejak et al, 2014;Bartolomei et al, 2008;David et al, 2011;Gnatkovsky et al, 2011Gnatkovsky et al, , 2014Varatharajah et al, 2018).…”

Section: Validation Of An Ez Biomarkermentioning

confidence: 99%

Learning to define an electrical biomarker of the epileptogenic zone

Grinenko

Mosher

et al. 2019

Human Brain Mapping

View full text Add to dashboard Cite

The role of fast activity as a potential biomarker in localization of the epileptogenic zone (EZ) remains controversial due to recently reported unsatisfactory performance.We recently identified a "fingerprint" of the EZ as a time-frequency pattern that is defined by a combination of preictal spike(s), fast oscillatory activity, and concurrent suppression of lower frequencies. Here we examine the generalizability of the fingerprint in application to an independent series of patients (11 seizure-free and 13 nonseizure-free after surgery) and show that the fingerprint can also be identified in seizures with lower frequency (such as beta) oscillatory activity. In the seizure-free group, only 5 of 47 identified EZ contacts were outside the resection. In contrast, in the non-seizure-free group, 104 of 142 identified EZ contacts were outside the resection. We integrated the fingerprint prediction with the subject's MR images, thus providing individualized anatomical estimates of the EZ. We show that these fingerprint-based estimates in seizure-free patients are almost always inside the resection. On the other hand, for a large fraction of the nonseizure-free patients the estimated EZ was not well localized and was partially or completely outside the resection, which may explain surgical failure in such cases. We also show that when mapping fast activity alone onto MR images, the EZ was often over-estimated, indicating a reduced discriminative ability for fast activity relative to the full fingerprint for localization of the EZ. K E Y W O R D Sepileptogenic zone, high-frequency oscillations, localization-related epilepsy, partial seizure, stereo-EEG

show abstract

“…Phase-amplitude coupling (PAC) is a form of cross-frequency coupling 43 where the amplitude of a higher frequency oscillation is modulated by the phase of a lower frequency oscillations. Recent studies [32][33][34][35] have shown that high PAC values are related with the SOZ. There are many proposed methods to estimate PAC [44][45][46][47][48][49][50] and different parameter choices that can be made (i.e.…”

Section: Phase Amplitude Couplingmentioning

confidence: 99%

“…There is no systematic study evaluating the performances of PAC in localizing the epileptogenic tissue using either different PAC implementations or the different parameters choices. We decided to investigate our dataset computing PAC between the modulating phase of theta band activity (4-8 Hz) and the amplitude of gamma activity (30-80 Hz) because this frequency band pairs were successfully investigated in recent epilepsy related studies [32][33][34][35] . For the detailed formula see Appendix .…”

Section: Phase Amplitude Couplingmentioning

confidence: 99%

See 1 more Smart Citation

To resect or not to resect? Unbiased performances of single and combined biomarkers in intra-operative corticography for tailoring during epilepsy surgery

Demuru

Kalitzin

Zweiphenning

et al. 2019

Preprint

View full text Add to dashboard Cite

Objective:Signal analysis biomarkers, in an intra-operative setting, may be complementary tools to guide and tailor the resection in drug-resistant epilepsy patients. Unbiased assessment of biomarker performances are needed to evaluate their clinical usefulness and translation. We defined a realistic ground-truth scenario and compared the effectiveness of different biomarkers alone and combined to localize epileptogenic tissue. Methods:We investigated the performances of univariate, bivariate and multivariate signal biomarkers applied to 1 minute inter-ictal intra-operative electrocorticography to discriminate between electrodes covering normal or pathologic activity in 47 drug-resistant people with epilepsy (temporal and extra-temporal) who had been seizure-free one year after the operation. Results:The best result using a single biomarker was obtained using the phase-amplitude coupling measure for which the epileptogenic tissue was localized in 16 out of 47 patients. Combining the whole set of biomarkers provided an improvement of the performances: 20 out of 47 patients. Repeating the analysis only on the temporal-lobe resections we reached a sensitivity of 93% (28 out of 30) combining all the biomarkers. Conclusion:We suggest that the assessment of biomarker performances on a ground-truth scenario is required to have a proper estimate on how biomarkers translate into clinical use. Phase-amplitude coupling seems the best performing single biomarker and combining biomarkers improves localization of epileptogenic tissue. However, sensitivity achieved is not adequate for the usage as a tool in the operation theater, but it can improve the understanding of pathophysiological process.

show abstract

Integrating artificial intelligence with real-time intracranial EEG monitoring to automate interictal identification of seizure onset zones in focal epilepsy

Cited by 68 publications

References 87 publications

NREM sleep is the state of vigilance that best identifies the epileptogenic zone in the interictal electroencephalogram

NREM sleep is the state of vigilance that best identifies the epileptogenic zone in the interictal electroencephalogram

Learning to define an electrical biomarker of the epileptogenic zone

To resect or not to resect? Unbiased performances of single and combined biomarkers in intra-operative corticography for tailoring during epilepsy surgery

Contact Info

Product

Resources

About