Tracey Camilleri | University of Malta (original) (raw)

Papers by Tracey Camilleri

Journal of Neural Engineering, 2018

OBJECTIVE Despite the vast research aimed at improving the performance of steady-state visually e... more OBJECTIVE Despite the vast research aimed at improving the performance of steady-state visually evoked potential (SSVEP)-based brain-computer interfaces (BCIs), several limitations exist that restrict the use of such applications for long-term users in the real-world. One of the main challenges has been to reduce training time while maintaining good BCI performance. In view of this challenge, this survey identifies and compares the different training requirements of feature extraction methods for SSVEP-based BCIs. APPROACH This paper reviews the various state-of-the-art SSVEP feature extraction methods that have been developed and are most widely used in the literature. MAIN RESULTS The main contributions compared to existing reviews are the following: (i) a detailed summary, including a brief mathematical description of each feature extraction algorithm, providing a guide to the basic concepts of the state-of-the-art techniques for SSVEP-based BCIs found in literature; (ii) a categorisation of the training requirements of SSVEP-based methods into three categories, defined as training-free methods, subject-specific and subject-independent training methods; (iii) a comparative review of the training requirements of SSVEP feature extraction methods, providing a reference for future work on SSVEP-based BCIs. SIGNIFICANCE This review highlights the strengths and weaknesses of the three categories of SSVEP training methods. Training-free systems are more practical but their performance is limited due to inter-subject variability resulting from the complex EEG activity. Feature extraction methods that incorporate some training data address this issue and in fact have outperformed training-free methods: subject-specific BCIs are tuned to the individual yielding the best performance at the cost of long, tiring training sessions making these methods unsuitable for everyday use; subject-independent BCIs that make use of training data from various subjects offer a good trade-off between training effort and performance, making these BCIs better suited for practical use.

Purpose: At six months post-stroke, 33-66% of survivors do not present with full recovery of uppe... more Purpose: At six months post-stroke, 33-66% of survivors do not present with full recovery of upper limb function. One of the methods of measuring cortical neurophysiological mechanisms of brain recovery in stroke, is electroencephalography (EEG). Disruption of neural connectivity can be measured by event related potentials such as somatosensory evoked potentials (SSEPs) and event related synchronisation. The aim of this systematic review was to examine the current evidence about the changes of cortical activity measured by EEG or magnetoencephalography (MEG) in association with sensorimotor upper limb impairments in stroke. Method: In order to identify the relevant studies, electronic searches, abstract and full-text papers were independently reviewed by two reviewers. From 1614 papers, 32 papers were selected for risk of bias assessment. Nine papers were then included in the review; 7 used EEG and 2 used MEG methodology. Results: In total, 321 people with stroke were included. Prel...

ACM Symposium on Eye Tracking Research and Applications, 2020

In this work, a novel method to estimate the ocular pose from electrooculography (EOG) signals is... more In this work, a novel method to estimate the ocular pose from electrooculography (EOG) signals is proposed. This method is based on an electrical battery model of the eye which relates the EOG potential to the distances between an electrode and the left/right cornea and retina centre points. In this work, this model is used to estimate the ocular angles (OAs), that is the orientation of the two ocular globes separately. Using this approach, an average cross-validated horizontal and vertical OA estimation error of 2.91 ± 0.86° and 2.42 ± 0.58° respectively was obtained. Furthermore, we show how these OA estimates may be used to estimate the gaze angles (GAs) without requiring the distance between the subject’s face-plane and the target-plane, as in previous work. Using the proposed method, a cross-validated horizontal and vertical GA estimation error of 2.13 ± 0.73° and 2.42 ± 0.58° respectively was obtained, which compares well with the previous distance-based GA estimation technique.

2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2019

In this work, a novel method to estimate the gaze angles using electrooculographic (EOG) signals ... more In this work, a novel method to estimate the gaze angles using electrooculographic (EOG) signals is presented. Specifically, this work investigates the use of a battery model of the eye, which relates the recorded EOG potential with the distances between the corresponding electrode and the centre points of the cornea and retina, for gaze angle estimation. Using this method a cross-validated horizontal and vertical gaze angle error of 2.42±0.91° and 2.30±0.50° respectively was obtained across six subjects, demonstrating that the proposed methods and the battery model may be used to estimate the user’s ocular pose reliably.

Sensors

Electroencephalogram (EEG)-based brain–computer interfaces (BCIs) provide a novel approach for co... more Electroencephalogram (EEG)-based brain–computer interfaces (BCIs) provide a novel approach for controlling external devices. BCI technologies can be important enabling technologies for people with severe mobility impairment. Endogenous paradigms, which depend on user-generated commands and do not need external stimuli, can provide intuitive control of external devices. This paper discusses BCIs to control various physical devices such as exoskeletons, wheelchairs, mobile robots, and robotic arms. These technologies must be able to navigate complex environments or execute fine motor movements. Brain control of these devices presents an intricate research problem that merges signal processing and classification techniques with control theory. In particular, obtaining strong classification performance for endogenous BCIs is challenging, and EEG decoder output signals can be unstable. These issues present myriad research questions that are discussed in this review paper. This review cov...

At six months post-stroke, 33-66% do not present with full recovery of upper limb function1. Iden... more At six months post-stroke, 33-66% do not present with full recovery of upper limb function1. Identified predictors for poor upper limb sensorimotor recovery are increased stroke severity, more severe somatosensory and motor impairments and the presence of visuospatial neglect2. Somatosensory deficits are experienced by 21-54% of stroke survivors and negatively impact on upper limb use, reaching, grasping and dexterity. One of the methods of measuring cortical neurophysiological mechanisms of brain recovery in stroke is using electroencephalography (EEG). Disruption of neural connectivity can be measured by Event Related Desynchronisation (ERD) using EEG. Decrease in ERD of alpha and beta-bands oscillations of the affected sensorimotor areas compared to contralesional areas have been reported at rest and during movement of people with stroke with motor impairments 3,4. Currently, there is a gap in knowledge about the understanding of the recovery of underlying cortical activity of pe...

2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2021

Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies, 2022

Sleep EEG data is characterised by various events that allow for the identification of the differ... more Sleep EEG data is characterised by various events that allow for the identification of the different sleep stages. Stage 2 in particular is characterised by two morphologically distinct waveforms, specifically spindles and K-complexes. Manual scoring of these events is time consuming and risks being subjectively interpreted; hence there is the need of robust automatic detection techniques. Various approaches have been adopted in the literature, ranging from period-amplitude analysis, to spectral analysis and autoregressive modelling. Most of the adopted techniques follow an episodic approach where the goal is to identify whether an epoch of EEG data contains an event, such as a spindle, or otherwise. The disadvantage of this approach is that it requires the data to be segmented into epochs, risking that an event falls at an epoch boundary, and it has low temporal resolution.

Biomedical Signal Processing and Control, 2020

Abstract Objective Electrooculography (EOG) is an eye movement recording technique based on the e... more Abstract Objective Electrooculography (EOG) is an eye movement recording technique based on the electrical activity due to the eyes, which may be used to develop human computer interfaces. The EOG signal baseline is subject to drifting and, although several baseline drift mitigation techniques have been proposed in the literature, the specific technique and the corresponding parameters are generally arbitrarily chosen. Furthermore, the literature does not establish which is the most suitable technique. Hence, this work aims to review these different techniques, and qualitatively and quantitatively compare their performance in mitigating the baseline drift using the same EOG data. This dataset is also being made publicly available to serve as a benchmark for future work. Methods The state-of-the-art baseline drift mitigation techniques, namely, frequent DC reference resetting, signal differencing, high-pass filtering, wavelet decomposition and polynomial fitting, were implemented and statistically compared. Results Generally, frequent resetting and signal differencing were statistically significantly better than the other techniques. Furthermore, high-pass filtering and wavelet decomposition had statistically similar performance, while the polynomial fitting technique was never superior to the other techniques. Conclusions While frequent resetting and signal differencing gave the best performance, the former disrupts the user's interaction with the system whereas the latter undesirably changes the EOG signal morphology. From the remaining techniques, high-pass filtering and wavelet decomposition would be the most suitable, but only the former would be applicable to real-time applications. Significance This work compares five state-of-the-art EOG baseline drift mitigation techniques and provides a guideline for future work.

Biomedical Signal Processing and Control, 2019

Biomedical Physics & Engineering Express, 2019

Brain-Computer Interfaces, 2018

Electronic Workshops in Computing, 2018

Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, Jul 1, 2017

Brain-computer interface (BCI) systems have emerged as an augmentative technology that can provid... more Brain-computer interface (BCI) systems have emerged as an augmentative technology that can provide a promising solution for individuals with motor dysfunctions and for the elderly who are experiencing muscle weakness. Steady-state visually evoked potentials (SSVEPs) are widely adopted in BCI systems due to their high speed and accuracy when compared to other BCI paradigms. In this paper, we apply combined magnitude and phase features for class discrimination in a real-time SSVEP-based BCI platform. In the proposed real-time system users gain control of a motorised bed system with seven motion commands and an idle state. Experimental results amongst eight participants demonstrate that the proposed real-time BCI system can successfully discriminate between different SSVEP signals achieving high information transfer rates (ITR) of 82.73 bits/min. The attractive features of the proposed system include noninvasive recording, simple electrode configuration, excellent BCI response and mini...

Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, Jul 1, 2017

The use of brain signals for person recognition has in recent years attracted considerable intere... more The use of brain signals for person recognition has in recent years attracted considerable interest because of the increased security and privacy these can offer when compared to conventional biometric measures. The main challenge lies in extracting features from the EEG signals that are sufficiently distinct across individuals while also being sufficiently consistent across multiple recording sessions. A range of EEG phenomena including eyes open and eyes closed activity, visual evoked potentials (VEPs) through image presentation, and other mental tasks have been studied for their use in biometry.

Biomedical Physics & Engineering Express, 2017

This work focusses on reducing the training time required for a brain–computer interface (BCI) mu... more This work focusses on reducing the training time required for a brain–computer interface (BCI) music player based on steady-state visually evoked potentials (SSVEPs). The music player is menu driven, featuring three different interfaces with up to six continuously flickering stimuli, similar to typical smart phone applications. This work investigates whether it is possible to go from a menu driven training approach to one which uses a single stimuli session only for training, or one which uses solely the data collected from the menu with the largest number of stimuli. Results show that the latter reduces the training time by 38.90%, specifically from 21 to 12.83 min without significant degradation in classification performance. Furthermore, promising results were also revealed when using a subject independent classifier which avoids individual training for new subjects by using training data from a database of other subjects. Although this work was targeted towards the brain controlled music player, the results are applicable to any SSVEP based BCI system having multiple interfaces with different number of flickering stimuli.

Journal of Neural Engineering, 2018

OBJECTIVE Despite the vast research aimed at improving the performance of steady-state visually e... more OBJECTIVE Despite the vast research aimed at improving the performance of steady-state visually evoked potential (SSVEP)-based brain-computer interfaces (BCIs), several limitations exist that restrict the use of such applications for long-term users in the real-world. One of the main challenges has been to reduce training time while maintaining good BCI performance. In view of this challenge, this survey identifies and compares the different training requirements of feature extraction methods for SSVEP-based BCIs. APPROACH This paper reviews the various state-of-the-art SSVEP feature extraction methods that have been developed and are most widely used in the literature. MAIN RESULTS The main contributions compared to existing reviews are the following: (i) a detailed summary, including a brief mathematical description of each feature extraction algorithm, providing a guide to the basic concepts of the state-of-the-art techniques for SSVEP-based BCIs found in literature; (ii) a categorisation of the training requirements of SSVEP-based methods into three categories, defined as training-free methods, subject-specific and subject-independent training methods; (iii) a comparative review of the training requirements of SSVEP feature extraction methods, providing a reference for future work on SSVEP-based BCIs. SIGNIFICANCE This review highlights the strengths and weaknesses of the three categories of SSVEP training methods. Training-free systems are more practical but their performance is limited due to inter-subject variability resulting from the complex EEG activity. Feature extraction methods that incorporate some training data address this issue and in fact have outperformed training-free methods: subject-specific BCIs are tuned to the individual yielding the best performance at the cost of long, tiring training sessions making these methods unsuitable for everyday use; subject-independent BCIs that make use of training data from various subjects offer a good trade-off between training effort and performance, making these BCIs better suited for practical use.

Purpose: At six months post-stroke, 33-66% of survivors do not present with full recovery of uppe... more Purpose: At six months post-stroke, 33-66% of survivors do not present with full recovery of upper limb function. One of the methods of measuring cortical neurophysiological mechanisms of brain recovery in stroke, is electroencephalography (EEG). Disruption of neural connectivity can be measured by event related potentials such as somatosensory evoked potentials (SSEPs) and event related synchronisation. The aim of this systematic review was to examine the current evidence about the changes of cortical activity measured by EEG or magnetoencephalography (MEG) in association with sensorimotor upper limb impairments in stroke. Method: In order to identify the relevant studies, electronic searches, abstract and full-text papers were independently reviewed by two reviewers. From 1614 papers, 32 papers were selected for risk of bias assessment. Nine papers were then included in the review; 7 used EEG and 2 used MEG methodology. Results: In total, 321 people with stroke were included. Prel...

ACM Symposium on Eye Tracking Research and Applications, 2020

In this work, a novel method to estimate the ocular pose from electrooculography (EOG) signals is... more In this work, a novel method to estimate the ocular pose from electrooculography (EOG) signals is proposed. This method is based on an electrical battery model of the eye which relates the EOG potential to the distances between an electrode and the left/right cornea and retina centre points. In this work, this model is used to estimate the ocular angles (OAs), that is the orientation of the two ocular globes separately. Using this approach, an average cross-validated horizontal and vertical OA estimation error of 2.91 ± 0.86° and 2.42 ± 0.58° respectively was obtained. Furthermore, we show how these OA estimates may be used to estimate the gaze angles (GAs) without requiring the distance between the subject’s face-plane and the target-plane, as in previous work. Using the proposed method, a cross-validated horizontal and vertical GA estimation error of 2.13 ± 0.73° and 2.42 ± 0.58° respectively was obtained, which compares well with the previous distance-based GA estimation technique.

2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2019

In this work, a novel method to estimate the gaze angles using electrooculographic (EOG) signals ... more In this work, a novel method to estimate the gaze angles using electrooculographic (EOG) signals is presented. Specifically, this work investigates the use of a battery model of the eye, which relates the recorded EOG potential with the distances between the corresponding electrode and the centre points of the cornea and retina, for gaze angle estimation. Using this method a cross-validated horizontal and vertical gaze angle error of 2.42±0.91° and 2.30±0.50° respectively was obtained across six subjects, demonstrating that the proposed methods and the battery model may be used to estimate the user’s ocular pose reliably.

Sensors

Electroencephalogram (EEG)-based brain–computer interfaces (BCIs) provide a novel approach for co... more Electroencephalogram (EEG)-based brain–computer interfaces (BCIs) provide a novel approach for controlling external devices. BCI technologies can be important enabling technologies for people with severe mobility impairment. Endogenous paradigms, which depend on user-generated commands and do not need external stimuli, can provide intuitive control of external devices. This paper discusses BCIs to control various physical devices such as exoskeletons, wheelchairs, mobile robots, and robotic arms. These technologies must be able to navigate complex environments or execute fine motor movements. Brain control of these devices presents an intricate research problem that merges signal processing and classification techniques with control theory. In particular, obtaining strong classification performance for endogenous BCIs is challenging, and EEG decoder output signals can be unstable. These issues present myriad research questions that are discussed in this review paper. This review cov...

At six months post-stroke, 33-66% do not present with full recovery of upper limb function1. Iden... more At six months post-stroke, 33-66% do not present with full recovery of upper limb function1. Identified predictors for poor upper limb sensorimotor recovery are increased stroke severity, more severe somatosensory and motor impairments and the presence of visuospatial neglect2. Somatosensory deficits are experienced by 21-54% of stroke survivors and negatively impact on upper limb use, reaching, grasping and dexterity. One of the methods of measuring cortical neurophysiological mechanisms of brain recovery in stroke is using electroencephalography (EEG). Disruption of neural connectivity can be measured by Event Related Desynchronisation (ERD) using EEG. Decrease in ERD of alpha and beta-bands oscillations of the affected sensorimotor areas compared to contralesional areas have been reported at rest and during movement of people with stroke with motor impairments 3,4. Currently, there is a gap in knowledge about the understanding of the recovery of underlying cortical activity of pe...

2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2021

Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies, 2022

Sleep EEG data is characterised by various events that allow for the identification of the differ... more Sleep EEG data is characterised by various events that allow for the identification of the different sleep stages. Stage 2 in particular is characterised by two morphologically distinct waveforms, specifically spindles and K-complexes. Manual scoring of these events is time consuming and risks being subjectively interpreted; hence there is the need of robust automatic detection techniques. Various approaches have been adopted in the literature, ranging from period-amplitude analysis, to spectral analysis and autoregressive modelling. Most of the adopted techniques follow an episodic approach where the goal is to identify whether an epoch of EEG data contains an event, such as a spindle, or otherwise. The disadvantage of this approach is that it requires the data to be segmented into epochs, risking that an event falls at an epoch boundary, and it has low temporal resolution.

Biomedical Signal Processing and Control, 2020

Abstract Objective Electrooculography (EOG) is an eye movement recording technique based on the e... more Abstract Objective Electrooculography (EOG) is an eye movement recording technique based on the electrical activity due to the eyes, which may be used to develop human computer interfaces. The EOG signal baseline is subject to drifting and, although several baseline drift mitigation techniques have been proposed in the literature, the specific technique and the corresponding parameters are generally arbitrarily chosen. Furthermore, the literature does not establish which is the most suitable technique. Hence, this work aims to review these different techniques, and qualitatively and quantitatively compare their performance in mitigating the baseline drift using the same EOG data. This dataset is also being made publicly available to serve as a benchmark for future work. Methods The state-of-the-art baseline drift mitigation techniques, namely, frequent DC reference resetting, signal differencing, high-pass filtering, wavelet decomposition and polynomial fitting, were implemented and statistically compared. Results Generally, frequent resetting and signal differencing were statistically significantly better than the other techniques. Furthermore, high-pass filtering and wavelet decomposition had statistically similar performance, while the polynomial fitting technique was never superior to the other techniques. Conclusions While frequent resetting and signal differencing gave the best performance, the former disrupts the user's interaction with the system whereas the latter undesirably changes the EOG signal morphology. From the remaining techniques, high-pass filtering and wavelet decomposition would be the most suitable, but only the former would be applicable to real-time applications. Significance This work compares five state-of-the-art EOG baseline drift mitigation techniques and provides a guideline for future work.

Biomedical Signal Processing and Control, 2019

Biomedical Physics & Engineering Express, 2019

Brain-Computer Interfaces, 2018

Electronic Workshops in Computing, 2018

Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, Jul 1, 2017

Brain-computer interface (BCI) systems have emerged as an augmentative technology that can provid... more Brain-computer interface (BCI) systems have emerged as an augmentative technology that can provide a promising solution for individuals with motor dysfunctions and for the elderly who are experiencing muscle weakness. Steady-state visually evoked potentials (SSVEPs) are widely adopted in BCI systems due to their high speed and accuracy when compared to other BCI paradigms. In this paper, we apply combined magnitude and phase features for class discrimination in a real-time SSVEP-based BCI platform. In the proposed real-time system users gain control of a motorised bed system with seven motion commands and an idle state. Experimental results amongst eight participants demonstrate that the proposed real-time BCI system can successfully discriminate between different SSVEP signals achieving high information transfer rates (ITR) of 82.73 bits/min. The attractive features of the proposed system include noninvasive recording, simple electrode configuration, excellent BCI response and mini...

Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, Jul 1, 2017

The use of brain signals for person recognition has in recent years attracted considerable intere... more The use of brain signals for person recognition has in recent years attracted considerable interest because of the increased security and privacy these can offer when compared to conventional biometric measures. The main challenge lies in extracting features from the EEG signals that are sufficiently distinct across individuals while also being sufficiently consistent across multiple recording sessions. A range of EEG phenomena including eyes open and eyes closed activity, visual evoked potentials (VEPs) through image presentation, and other mental tasks have been studied for their use in biometry.

Biomedical Physics & Engineering Express, 2017

This work focusses on reducing the training time required for a brain–computer interface (BCI) mu... more This work focusses on reducing the training time required for a brain–computer interface (BCI) music player based on steady-state visually evoked potentials (SSVEPs). The music player is menu driven, featuring three different interfaces with up to six continuously flickering stimuli, similar to typical smart phone applications. This work investigates whether it is possible to go from a menu driven training approach to one which uses a single stimuli session only for training, or one which uses solely the data collected from the menu with the largest number of stimuli. Results show that the latter reduces the training time by 38.90%, specifically from 21 to 12.83 min without significant degradation in classification performance. Furthermore, promising results were also revealed when using a subject independent classifier which avoids individual training for new subjects by using training data from a database of other subjects. Although this work was targeted towards the brain controlled music player, the results are applicable to any SSVEP based BCI system having multiple interfaces with different number of flickering stimuli.