Agency for Science, Technology and Research

Tan Tock Seng Hospital Pte Ltd

National University of Singapore

Kent Ridge Digital Labs

National Healthcare Group Pte Ltd

Institute of Mental Health

Cuntai Guan

Haihong Zhang

Kai Keng Ang

Kok Soon Phua

Hai Hong Zhang

Chuan Chu Wang

Zheng Yang Chin

Chuanchu Wang

Zhuo Zhang

Huijuan Yang

Keng Peng Tee

Tih Shih Lee

Yaozhang Pan

Juanhong Yu

Ranga Krishnan

Brahim Ahmed Salah Hamadi Charef

Jun Xu

Haizhou Li

Soon Huat Ng

Ti Eu Chan

Jit Biswas

Shijian Lu

Choon Guan Lim

Chu Chuan Wang

Aung Aung Phyo Wai

Su-Yin Yang

Jayachandran Maniyeri

Fong Ling Loy

Ka Yin Christina Tang

Yongwei Zhu

Rosa Qi Yue So

Siang Fook Foo

Keong Andrew Ng

Emily Jian Zhong Hao

Zhou Zhang

Soon Yin Tjan

Jiliang Eugene Phua

Sui Geok Karen Chua

Longjiang Zhou

Hock Guan Adrian Tan

Brahim Hamadi Charef

Daniel Fung

Brahim Ahmed Hamadi Charef

Disclosed is a process for identifying and extracting pain-related electroencephalogram (EEG) signals. The process comprises receiving, from one or more trials, EEG data for each trial; determining a current density for each signal; estimating the current density for a set of neural activity regions of interest, based on the computed current density; and computing at least one spectrum characteristic for each trial based on the estimated current density. Thus mean and variance of changes in the EEG data between EEG data labeled as being indicative of a pain state and EEG data labeled as being indicative of a non-pain state, for each neural activity region of interest can be calculated, and pain-related EEG signals can be identified based on at least one a region of interest at which the variance is below a predetermined threshold.

Identifying and extracting electroencephalogram signals

Disclosed is a signal processing system for a brain machine interface (BMI) and a method performed by such a signal processing system. The method involves receiving and filtering input data comprising cortical signal samples, calculating a threshold for detection of spikes, identifying spikes in subsequently received samples, and outputting spike data corresponding to the identified spike or spikes.

Signal processing system for brain machine interface and method performed by the system

Described is a computer system for establishing an electroencephalogram (EEG) model for discriminating between alert and fatigue states. The computer system comprises a receiver module for receiving an alert state segment illustrative of an alert state of at least one subject, and one or more EEG fatigue data segments illustrative of a fatigue state of the at least one subject. The computer system further comprises a segment selector for selecting one of the one or more fatigue data segments and setting it to be an assumed maximum fatigue segment, an EEG classifier trainer for training an EEG classifier by extracting an EEG feature space from the alert state segment and assumed maximum fatigue segment, and a maximum fatigue identifier module for identifying a segment of maximum fatigue by applying the EEG classifier to each of the fatigue data segments. The computer system further comprises a segment comparator for determining if the segment of maximum fatigue is consistent with the assumed maximum fatigue segment, and a limit setter for setting the segment of maximum fatigue as a revised assumed maximum fatigue segment, if the segment of maximum fatigue is inconsistent with the assumed maximum fatigue segment, and supplying the EEG classifier trainer with the revised assumed maximum fatigue segment. The computer system further comprises a model output module for setting the EEG classifier as the EEG model for discriminating between alert and fatigue states in segments of EEG data, if the segment of maximum fatigue is consistent with the assumed maximum fatigue segment.

Alert versus fatigue discriminator

A system for detecting eye blink activity, the system including one or more processors in communication with non-transitory data storage media having instructions stored thereon that, when executed by the one or more processors configure the one or more processors to perform the steps of receiving, with a receiving module, a signal associated with eye blink activity; filtering, at a processing module, the signal to reduce noise from the signal; calculating, at the processing module, a variance of the filtered signal for estimating background noise; removing, at the processing module, background noise associated with the variance; and detecting, at the processing module, eye blink activity by performing duration detection.

System and method for detecting eye activity

A method for profiling sleep of an individual is provided. The method includes defining a sleep feature space for the individual, measuring a brain wave for the individual during the individual's sleep, and mapping the sleep feature space in response to a comparison of the brain wave and a previous brain wave measurement used to define the sleep feature space. The brain wave may comprise a brain wave spectrum. The sleep feature space may comprise, or be composed of, spectral power and envelope measures. The method also includes modelling the mapped sleep feature space in response to recognized neural network patterns corresponding to each of a plurality of sleep stages derived from recognizing the neural network patterns from the sleep feature space and deriving a sleep profile for the individual from sleep stages determined in response to the modelled mapped sleep feature space and the brain wave of the individual.

Sleep profiling system with feature generation and auto-mapping

A computer system for acquiring a control command. The system includes one or more processors in communication with non-transitory data storage media having instructions stored thereon that, when executed by the one or more processors, configure the one or more processors to perform particular steps. The steps include: receiving, at an electroencephalogram (EEG) receiver, baseline EEG data; and acquiring, using a normalisation module, one or more normalisation factors from the baseline EEG data. The steps further include: receiving, at the EEG receiver, stimulated EEG data sensed by one or more EEG sensors; and correlating the stimulated EEG data across the one or more EEG sensors, using a correlation module, to generate one or more correlation coefficients corresponding to one or more stimulation frequencies, each of the one or more stimulation frequencies corresponding to a respective candidate control command.

Computer system for acquiring a control command

A method for calibrating and executing a rehabilitation exercise for a stroke-affected limb of a stroke patient is disclosed, the method comprising the steps of providing a haptic device for an able limb of the stroke patient to manipulate to perform a calibration action to result in a first position of the haptic device, and providing the haptic device for the stroke-affected limb to manipulate to perform the calibration action to result in a second position of the haptic device. The method further comprises the steps of moving the haptic device coupled with the stroke-affected limb from the second position towards the first position until a predetermined counterforce is detected, indicating an extreme position for the stroke-affected limb using the haptic device, and calibrating the haptic device with the extreme position such that during the rehabilitation exercise, the haptic device is prevented from moving beyond the extreme position.

Method and system for using haptic device and brain-computer interface for rehabilitation

A method for ballistocardiography (BCG) and a BCG system ballistocardiography (BCG) system for heart beat determination are provided. The method includes digitizing a plurality of signals received from a corresponding plurality of sensors, estimating a plurality of smoothed cepstra corresponding to each of the plurality of digitized signals in response to a smoothed cepstrum analysis of a digital signal at a reception time of each of the plurality of digitized signals, and estimating a fused cepstrum for the plurality of digitized signals in response to the plurality of smoothed cepstra. The method further includes determining a heart rate in response to the plurality of smoothed cepstra and the fused cepstrum.

Multi-channel ballistocardiography with cepstrum smoothing and quality-based dynamic channel selection

A cognitive training method has a step of obtaining sensor data of a subject during a memory exercise which in turn determines whether the sensor data includes predetermined information indicative of cognitive functions used for remembering. The step of advancing the memory exercise is carried out when it is determined that the sensor data includes predetermined information indicative of cognitive functions used for remembering. The step of calculating an objective cognitive assessment of the subject is carried out when the memory exercise is completed and cognitive training is measured at least partially in response to the advancing of the memory exercise. The fact that an objective cognitive assessment is calculated provides a method of using an autodidactic cognitive training device.

Autodidactic cognitive training device and method thereof

A method and system for sound-induced sleep is provided. The method includes detecting brainwave signals of a subject. The method further includes analyzing the brainwave signals to determine a current sleepy state of the subject and selecting a sound in response to the current sleepy state based on a quantified association index between the brainwave signals and acoustical features of the sound. The method further more includes playing the sound to the subject.

Sound-induced sleep method and a system therefor

A method for deriving optimal discriminating features indicative of a subject state when the subject listens to one of a set of musical pieces, comprising a step of extracting frequency features from the subject's EEG signal when the subject is in a first subject state and a second subject state, the frequency features being extracted from more than one frequency band in one set of time segments; and identifying optimal discriminating features from the extracted frequency features, the optimal discriminating features indicative of characteristics of the EEG signal when the subject is in the first subject state and the second subject state, wherein one of the first subject state and the second subject state indicates that the subject likes a musical piece while the other state indicates that the subject does not like the musical piece.

System and method for developing a model indicative of a subject's emotional state when listening to musical pieces

A method of training a classification algorithm for a Brain Computer Interface (BCI). The method includes the steps of: dividing a Electroencephalography (EEG) signal into a plurality of time segments; for each time segment, dividing a corresponding EEG signal portion into a plurality of frequency bands; for each frequency band, computing a spatial filtering projection matrix based on a Common Spatial Pattern (CSP) algorithm and a corresponding feature, and computing mutual information of each corresponding feature with respect to one or more motor imagery classes; for each time segment, summing the mutual information of all the corresponding features with respect to the respective classes; and selecting the corresponding features of the time segment with a maximum sum of mutual information for one class for training classifiers of the classification algorithm.

Brain-computer interface system and method

A method of calibrating a motor imagery detection module and a system for motor rehabilitation are provided. The method comprises acquiring Electroencephalography (EEG) data from a subject; selecting classification features from the EEG data; wherein the feature selection comprises modelling an idle state &#x3c9;of the subject by M sub-classes &#x3c7;, j=1, . . . , M.

Method and system for motor rehabilitation

A device and method for generating a representation of a subject's attention level. The device measures brain signals from the subject; extracts temporal features from the brain signals; classifies the extracted temporal features using a classifier to give a score X; extracts spectral-spatial features from the brain signals; selects spectral-spatial features containing discriminative information between concentration and non-concentration states from the set of extracted spectral-spatial features; classifies the selected spectral-spatial features using a classifier to give a score X; combines the scores Xand Xto give a single score; and presents the score to the subject.

Device and method for generating a representation of a subject's attention level

A method and system for motor learning. The method comprises the steps of detecting a user's motor intent; and giving robotic assistance to the user for executing a motor task associated with the motor intent based on the detected motor intent.

System and method for motor learning

Growing community of inventors

Singapore, Singapore

Cuntai Guan