Semi-Supervised Anomaly Detection for EEG Waveforms Using Deep Belief Nets

Wulsin, Drausin; Blanco, Justin A.; Mani, Ram S.; Litt, Brian

doi:10.1109/icmla.2010.71

Cited by 74 publications

(47 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These classes are very similar to what others have used [4] to perform stroke and epilepsy detection. In fact, we have replicated state of the art results on these tasks using the technology described in this paper.…”

Section: Preliminary Experimentsmentioning

confidence: 68%

“…The highest F score reported in [4] is 0.476 while our system produced an F score of 0.702 for the evaluation data and 0.772 for the training data on the TUH EEG Corpus. Figure 6 shows a tradeoff between false alarms and detections (correct recognitions).…”

Section: Preliminary Experimentsmentioning

confidence: 96%

“…To put these results in perspective, we have also compared our system to Wulsin's approach using Deep Belief Networks [4]. Wulsin et al used a small proprietary database of clinical EEGs that was manually transcribed by two neurologists.…”

Section: Preliminary Experimentsmentioning

confidence: 99%

See 2 more Smart Citations

The TUH EEG CORPUS: A big data resource for automated EEG interpretation

Harati

Lopez

Obeid

et al. 2014

2014 IEEE Signal Processing in Medicine and Biology Symposium (SPMB)

View full text Add to dashboard Cite

The Neural Engineering Data Consortium (NEDC) is releasing its first major big data corpus -the Temple University Hospital EEG Corpus. This corpus consists of over 25,000 EEG studies, and includes a neurologist's interpretation of the test, a brief patient medical history and demographic information about the patient such as gender and age. For the first time, there is a sufficient amount of data to support the application of state of the art machine learning algorithms. In this paper, we present pilot results of experiments on the prediction of some basic attributes of an EEG from the raw EEG signal data using a 3,762 session subset of the corpus. Standard machine learning approaches are shown to be capable of predicting commonly occurring events from simple features with high accuracy on closed-loop testing, and can deliver error rates below 50% on a 6-way open set classification problem. This is very promising performance since commercial technology fails on this data.

show abstract

Section: Preliminary Experimentsmentioning

confidence: 68%

Section: Preliminary Experimentsmentioning

confidence: 96%

See 1 more Smart Citation

The TUH EEG CORPUS: A big data resource for automated EEG interpretation

Harati

Lopez

Obeid

et al. 2014

2014 IEEE Signal Processing in Medicine and Biology Symposium (SPMB)

View full text Add to dashboard Cite

show abstract

“…Deep learning has been applied to many domains, such as biomedical signals EEG Wulsin et al, 2010;Wulsin et al, 2011), EMG, EOG (Wang and Shang, 2013), and EEG, EMG and EOG (Langkvist et al, 2012). These studies showed that deep learning can be applied to raw physiological data to learn relevant features effectively.…”

Section: Introductionmentioning

confidence: 98%

Automatic sleep stage classification based on sparse deep belief net and combination of multiple classifiers

Zhang

Bai

et al. 2015

Transactions of the Institute of Measurement and Control

View full text Add to dashboard Cite

This paper presents an automatic sleep stage method combining a sparse deep belief net and combination of multiple classifiers for electroencephalogram, electrooculogram and electromyogram. The sparse deep belief net was applied to extract features from these signals automatically, and the combination of multiple classifiers, utilizing the extracted features, assigned each 30-s epoch to one of the five possible sleep stages. More importantly, we proposed a new voting principle based on classification entropy to enhance the classification performance further by harnessing the complementary information provided by the individual classifier. Differently from existing methods, our method used unsupervised feature learning to extract features automatically from raw sleep data and classification based on the learned features. The results of automatic and manual scorings were compared on an epoch-by-epoch basis. The accuracies for wake, S1, S2, SWS and REM were 98.49%, 80.05%, 91.2%, 98.22% and 95.31%, respectively, and the total accuracy of sleep stage was 91.31%. The results demonstrated that the sparse deep belief net was an efficient feature extraction method for sleep data, and the combination of multiple classifiers based on classification entropy performed well on sleep stages.

show abstract

“…This can focus on identifying anomalous sequences with respect to a database of normal sequences, identifying an anomalous subsequence within a long sequence, or identifying a pattern in a sequence whose frequency of occurrence is anomalous [7]. Applications of anomaly detection are widespread and vary from intrusion detection in computer networks [31], credit card fraud detection [30], medical applications such as EEG analysis [29], to forensic applications such as the detection of abnormal behavior in surveillance videos [2]. Major challenges consist of sieving out the usually infrequently occurring anomalies from the total (often massive) data, and handling the problem of adaptation: certain (but not all) events that primarily occur as anomalies become accepted as 'normal' events, and should not be detected over and over again.…”

Section: Anomaly-detection Methodsmentioning

confidence: 99%

Anomaly detection for internet surveillance

et al. 2012

View full text Add to dashboard Cite

Many threats in the real world can be related to activity of persons on the internet. Internet surveillance aims to predict and prevent attacks and to assist in finding suspects based on information from the web. However, the amount of data on the internet rapidly increases and it is time consuming to monitor many websites. In this paper, we present a novel method to automatically monitor trends and find anomalies on the internet. The system was tested on Twitter data. The results showed that it can successfully recognize abnormal changes in activity or emotion.

show abstract

Semi-Supervised Anomaly Detection for EEG Waveforms Using Deep Belief Nets

Cited by 74 publications

References 6 publications

The TUH EEG CORPUS: A big data resource for automated EEG interpretation

The TUH EEG CORPUS: A big data resource for automated EEG interpretation

Automatic sleep stage classification based on sparse deep belief net and combination of multiple classifiers

Anomaly detection for internet surveillance

Contact Info

Product

Resources

About