Disclosing hidden information in the electroencephalogram using advanced signal analytical techniques

Lüttjohann, Annika

doi:10.1113/jp275132

Cited by 3 publications

(4 citation statements)

References 6 publications

(12 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using the output of the algorithm, we calculated the event-triggered average: (ie, the average of all segments of one signal aligned to the time of events in another signal) of 50 random examples of strong SWDs (distance to support vector of >1 to 2), moderate SWDs (>0.5 to 1), weak SWDs (>0 to 0.5), weak nonSWDs (<0 to −0.5), moderate nonSWDs (<−0.5 to −1), and strong non-SWDs (<−1 to −2), and 50 random locations not identified by our algorithm to delta (0.5-4 Hz), theta (6-9 Hz), sigma (10)(11)(12)(13)(14), and gamma (25-100 Hz) power surrounding each event during lights-on and lights-off periods. Using the output of the algorithm, we calculated the event-triggered average: (ie, the average of all segments of one signal aligned to the time of events in another signal) of 50 random examples of strong SWDs (distance to support vector of >1 to 2), moderate SWDs (>0.5 to 1), weak SWDs (>0 to 0.5), weak nonSWDs (<0 to −0.5), moderate nonSWDs (<−0.5 to −1), and strong non-SWDs (<−1 to −2), and 50 random locations not identified by our algorithm to delta (0.5-4 Hz), theta (6-9 Hz), sigma (10)(11)(12)(13)(14), and gamma (25-100 Hz) power surrounding each event during lights-on and lights-off periods.…”

Section: Application Of the Algorithm And Preliminary Testing Of Thmentioning

confidence: 95%

“…First, the definition of "epileptiform" is often vague [10][11][12] and is subject to disagreement. In addition, many electrographic features appear hidden in the time domain 14 and thus may be difficult to detect via visual inspection of EEG records. In addition, many electrographic features appear hidden in the time domain 14 and thus may be difficult to detect via visual inspection of EEG records.…”

Section: Introductionmentioning

confidence: 99%

“…Clinicians and researchers typically use subjective criteria, such as that an event "stands out of the background," 13 to force events into binary categories. In addition, many electrographic features appear hidden in the time domain 14 and thus may be difficult to detect via visual inspection of EEG records. As we show, events can appear ambiguous to the same expert scorer such that, on repeated presentations, a scorer changes their mind about an event's categorization.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

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

2019

View full text Add to dashboard Cite

Summary Objective Manual 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. Methods We 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. Results The 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. Significance Although 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).

show abstract

Section: Application Of the Algorithm And Preliminary Testing Of Thmentioning

confidence: 95%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

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

2019

View full text Add to dashboard Cite

show abstract

“…Clinicians and researchers typically use subjective criteria, such as that an event 'stands out of the background' 13 , to force events into binary categories. Additionally, many electrographic features appear hidden in the time domain 14 and thus may be difficult to detect via visual inspection of EEG records. As we show, events can appear ambiguous to the same expert scorer such that, on repeated presentations, a scorer changes their mind about an event's categorization.…”

Section: Introductionmentioning

confidence: 99%

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

Pfammatter

Maganti

Jones

2018

Preprint

View full text Add to dashboard Cite

and KeywordsObjective. Manual detection of spike-wave discharges (SWDs) from EEG records is time intensive, costly, and subject to inconsistencies/biases. Additionally, 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 pre-selected events to establish a confidence-based, continuous-valued scoring.Methods. We develop a support vector machine (SVM)-based algorithm for the automated detection of SWDs in the γ 2R43Q 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 a SVM to classify (SWD/nonSWD) andassign confidence scores to each event identified from 60 24-hour EEG records. We provide a detailed assessment of intra-and inter-rater scoring that demonstrates advantages of automated scoring.Results. The 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.Significance. While 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 (i.e., the most reliable features agreed upon by humans that also correlate with objective physiological measures). Key Point Box• Clinicians and researchers may benefit from an automated method of SWD detection that provides a framework for the quantitative description of SWDs and how they relate to other electrographic events.• We present an algorithm for the automated, consistent, and rapid scoring of SWDs that assigns a confidence to detected events that is highly correlated with human scoring confidence.• We characterize the human inter-and intra-rater consistency in the scoring of potential SWD events and compare them with the algorithm.• Events along the scoring continuum generated by the algorithm are temporally and proportionately correlated with changes in spectral power bands relevant to behavioral states including sleep.

show abstract

Diagnostic principles of immature epileptic (proepileptic) EEG activity in rats with genetic predisposition to absence epilepsy

Sitnikova

Smirnov

Grubov

et al. 2019

Informacionno-upravlâûŝie sistemy

View full text Add to dashboard Cite

Introduction:Absence epilepsy is a specific neurological disorder characterized by brief episodes of loss of consciousness (absence) accompanied by high-amplitude “spike-wave” discharges in the electroencephalogram (EEG). WAG/Rij rats with a genetic predisposition to absence epilepsy are used as a reliable model of this disease. This model is beneficial for investigating basic mechanisms of absence epilepsy, including the development of spike-wave seizures.Purpose:Establishing diagnostic principles for immature forms of spikewave activity in EEG (so-called proepileptic activity) of WAG/Rij rats.Results:Diagnostic criteria are proposed for proepileptic EEG activity in rats, based on time-frequency analysis with the continuous wavelet transform and skeletons of wavelet surfaces. The algorithm was tested in “epileptic” and “asymptomatic” individuals. Rats with the “epileptic” phenotype demonstrated a decrease in number of proepileptic patterns between 5 and 7 months of age in parallel to an increase in the number of epileptic discharges that might relate to the transformation of proepileptic activity to epileptic. In rats with the “asymptomatic” phenotype, the age-related decline in the number of proepileptic patterns was not accompanied by the occurrence of spike-wave discharges. A decrease in the instantaneous frequency was found in a larger number of proepileptic patterns in “epileptic” WAG/Rij rats as compared to “asymptomatic” individuals.Practical relevance:A similar approach can be used for early diagnosis of absence epilepsy at the preclinical stage in patients with genetic predisposition.

show abstract

Disclosing hidden information in the electroencephalogram using advanced signal analytical techniques

Cited by 3 publications

References 6 publications

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

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

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

Diagnostic principles of immature epileptic (proepileptic) EEG activity in rats with genetic predisposition to absence epilepsy

Contact Info

Product

Resources

About