Dataset Information

An automated, machine learning-based detection algorithm for spike-wave discharges (SWDs) in a mouse model of absence epilepsy

ABSTRACT: Manual detection of spike-wave discharges (SWDs) from EEG records is time intensive, costly and subject to inconsistencies and biases. Additionally, manual scoring often omits information on SWD confidence/intensity which may be important for the investigation of mechanistic-based research questions. While there are some automated and semi-automated methods for the detection of SWDs in humans and rats, there has been minimal development of these methods for SWDs in mice. Here we develop a support vector machine (SVM)-based algorithm for the automated detection of SWDs in the gamma2R43Q mouse model of absence epilepsy. The algorithm first identifies putative SWD events using frequency- and amplitude-based peak detection. Four humans experienced at identifying SWDs scored a set of 2500 putative events identified by the algorithm. Then, using predictors calculated from the Morlet wavelet transform of each event and the labels from the human scored events, we trained a SVM to classify (SWD/nonSWD) and assign confidence scores to each event identified from 60 24-hour EEG records (11 animals). We provide a detailed assessment of intra- and inter-rater scoring that demonstrate the advantages in reliability of automated scoring over human scoring. The result of this algorithm is a scoring of SWDs along a confidence continuum that is highly correlated with human confidence and that allows us to more effectively characterize ambiguous events in order to gain a better understanding of ground truth in SWD detection (i.e., the most reliable features agreed upon by human scorers that also correlate with objective physiological measures). Finally, we demonstrate that events along our scoring continuum are temporally and proportionately correlated, on the scale of seconds, with abrupt changes in spectral power bands relevant to normal behavioral states including sleep. Our results demonstrate the value of viewing detection of SWDs as a continuous, rather than discrete, classification problem.

SUBMITTER: Jesse Pfammatter

PROVIDER: S-BSST208 | biostudies-other |

REPOSITORIES: biostudies-other

ACCESS DATA

Similar Datasets

Project description:ObjectiveManual detection of spike-wave discharges (SWDs) from electroencephalography (EEG) records is time intensive, costly, and subject to inconsistencies/biases. In addition, manual scoring often omits information on SWD confidence/intensity, which may be important for the investigation of mechanistic-based research questions. Our objective is to develop an automated method for the detection of SWDs in a mouse model of absence epilepsy that is focused on the characteristics of human scoring of preselected events to establish a confidence-based, continuous-valued scoring.MethodsWe develop a support vector machine (SVM)-based algorithm for the automated detection of SWDs in the γ2R43Q mouse model of absence epilepsy. The algorithm first identifies putative SWD events using frequency- and amplitude-based peak detection. Four humans scored a set of 2500 putative events identified by the algorithm. Then, using predictors calculated from the wavelet transform of each event and the labels from human scoring, we trained an SVM to classify (SWD/nonSWD) and assign confidence scores to each event identified from 60, 24-hour EEG records. We provide a detailed assessment of intra- and interrater scoring that demonstrates advantages of automated scoring.ResultsThe algorithm scored SWDs along a continuum that is highly correlated with human confidence and that allows us to more effectively characterize ambiguous events. We demonstrate that events along our scoring continuum are temporally and proportionately correlated with abrupt changes in spectral power bands relevant to normal behavioral states including sleep.SignificanceAlthough there are automated and semi-automated methods for the detection of SWDs in humans and rats, we contribute to the need for continued development of SWD detection in mice. Our results demonstrate the value of viewing detection of SWDs as a continuous classification problem to better understand "ground truth" in SWD detection (ie, the most reliable features agreed upon by humans that also correlate with objective physiologic measures).

Project description:IntroductionDifferent studies have investigated by means of EEG-fMRI coregistration the brain networks related to generalized spike-and-wave discharges (GSWD) in patients with idiopathic generalized epilepsy (IGE). These studies revealed a widespread GSWD-related neural network that involves the thalamus and regions of the default mode network. In this study we investigated which brain regions are critically involved in the termination of absence seizures (AS) in a group of IGE patients.MethodsEighteen patients (6 male; mean age 25 years) with AS were included in the EEG-fMRI study. Functional data were acquired at 3T with continuous simultaneous video-EEG recording. Event-related analysis was performed with SPM8 software, using the following regressors: (1) GSWD onset and duration; (2) GSWD offset. Data were analyzed at single-subject and at group level with a second level random effect analysis.ResultsA mean of 17 events for patient was recorded (mean duration of 4.2 sec). Group-level analysis related to GSWD onset respect to rest confirmed previous findings revealing thalamic activation and a precuneus/posterior cingulate deactivation. At GSWD termination we observed a decrease in BOLD signal over the bilateral dorsolateral frontal cortex respect to the baseline (and respect to GSWD onset). The contrast GSWD offset versus onset showed a BOLD signal increase over the precuneus-posterior cingulate region bilaterally. Parametric correlations between electro-clinical variables and BOLD signal at GSWD offset did not reveal significant effects.ConclusionThe role of the decreased neural activity of lateral prefrontal cortex at GSWD termination deserve future investigations to ascertain if it has a role in promoting the discharge offset, as well as in the determination of the cognitive deficits often present in patients with AS. The increased BOLD signal at precuneal/posterior cingulate cortex might reflect the recovery of neural activity in regions that are "suspended" during spike and waves activity, as previously hypothesized.

Project description:Generalized spike-wave discharges in idiopathic generalized epilepsy are conventionally assumed to have abrupt onset and offset. However, in rodent models, discharges emerge during a dynamic evolution of brain network states, extending several seconds before and after the discharge. In human idiopathic generalized epilepsy, simultaneous EEG and functional MRI shows cortical regions may be active before discharges, and network connectivity around discharges may not be normal. Here, in human idiopathic generalized epilepsy, we investigated whether generalized spike-wave discharges emerge during a dynamic evolution of brain network states. Using EEG-functional MRI, we studied 43 patients and 34 healthy control subjects. We obtained 95 discharges from 20 patients. We compared data from patients with discharges with data from patients without discharges and healthy controls. Changes in MRI (blood oxygenation level-dependent) signal amplitude in discharge epochs were observed only at and after EEG onset, involving a sequence of parietal and frontal cortical regions then thalamus (P < 0.01, across all regions and measurement time points). Examining MRI signal phase synchrony as a measure of functional connectivity between each pair of 90 brain regions, we found significant connections (P < 0.01, across all connections and measurement time points) involving frontal, parietal and occipital cortex during discharges, and for 20 s after EEG offset. This network prominent during discharges showed significantly low synchrony (below 99% confidence interval for synchrony in this network in non-discharge epochs in patients) from 16 s to 10 s before discharges, then ramped up steeply to a significantly high level of synchrony 2 s before discharge onset. Significant connections were seen in a sensorimotor network in the minute before discharge onset. This network also showed elevated synchrony in patients without discharges compared to healthy controls (P = 0.004). During 6 s prior to discharges, additional significant connections to this sensorimotor network were observed, involving prefrontal and precuneus regions. In healthy subjects, significant connections involved a posterior cortical network. In patients with discharges, this posterior network showed significantly low synchrony during the minute prior to discharge onset. In patients without discharges, this network showed the same level of synchrony as in healthy controls. Our findings suggest persistently high sensorimotor network synchrony, coupled with transiently (at least 1 min) low posterior network synchrony, may be a state predisposing to generalized spike-wave discharge onset. Our findings also show that EEG onset and associated MRI signal amplitude change is embedded in a considerably longer period of evolving brain network states before and after discharge events.

Dataset Information

An automated, machine learning-based detection algorithm for spike-wave discharges (SWDs) in a mouse model of absence epilepsy

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets