Nikesh Bajaj
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Nikesh Bajaj has a total of 64 co-authors. Here are the top co-authors by number of shared publications:
Francesco Bellotti
4 shared publicationsAlessandro De Gloria
3 shared publications- Julie Wall
School of Computing and Engineering, University of West London, London W5 5RF, UK
3 shared publications - Mansour Moniri
University of East London
3 shared publications - Pratheep Kumar Paranthaman
Elon University
3 shared publications Riccardo Berta
3 shared publications
Latest Publications
- Open Access
Electroencephalogram-Based Analysis of Monomodal and Multimodal Interaction in Mixed Reality Games
Mixed reality (MR) technologies enable users to experience computer-generated content within the physical environment through spatial computing and head-mounted displays. By supporting real-time interaction through speech, gesture, gaze, and movement, MR offers new opportunities for game design beyond productivity and educational applications. However, relatively few studies have examined interaction modalities in MR games. In this paper, we present the design and deployment of four MR games on the Microsoft HoloLens 2: three that use monomodal input (speech, gaze, or gesture) and one that uses multimodal input (speech, gaze, and gesture). We conducted a study with ten participants and evaluated player experience using subjective self-reports of task load, emotional engagement, and comfort alongside objective measures, namely brain activity data collected with a five-channel electroencephalogram (EEG) device. Our preliminary findings suggest two clusters of interaction modalities based on subjective measures, a pattern that is also reflected in the objective EEG measures. Our analysis combining subjective and EEG data indicates that interaction modality influences task load and emotional engagement. Additionally, our functional connectivity analysis showed links in activity across the prefrontal, temporal, and occipital brain regions for different input modalities in the MR games.
Open AccessElectroencephalogram-Based Analysis of Monomodal and Multimodal Interaction in Mixed Reality Games
Mixed reality (MR) technologies enable users to experience computer-generated content within the physical environment through spatial computing and head-mounted displays. By supporting real-time interaction through speech, gesture, gaze, and movement, MR offers new opportunities for game design beyond productivity and educational applications. However, relatively few studies have examined interaction modalities in MR games. In this paper, we present the design and deployment of four MR games on the Microsoft HoloLens 2: three that use monomodal input (speech, gaze, or gesture) and one that uses multimodal input (speech, gaze, and gesture). We conducted a study with ten participants and evaluated player experience using subjective self-reports of task load, emotional engagement, and comfort alongside objective measures, namely brain activity data collected with a five-channel electroencephalogram (EEG) device. Our preliminary findings suggest two clusters of interaction modalities based on subjective measures, a pattern that is also reflected in the objective EEG measures. Our analysis combining subjective and EEG data indicates that interaction modality influences task load and emotional engagement. Additionally, our functional connectivity analysis showed links in activity across the prefrontal, temporal, and occipital brain regions for different input modalities in the MR games.
- 2 January 2026
- Published by Springer Nature
- p. 15-23
Exploring Immersive Gaming Experiences: A Comparative Analysis of Game Platform Features
- 2 January 2026
- Published by Springer Nature
- p. 15-23
Exploring Immersive Gaming Experiences: A Comparative Analysis of Game Platform Features
- Open Access
- 20 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 118-124
Self-supervise Learning for Medical Image Segmentation Using Barlow Twins Model
- 20 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 118-124
Open AccessSelf-supervise Learning for Medical Image Segmentation Using Barlow Twins Model
- Open Access
- 1 December 2024
- Published by Wolters Kluwer Health in Circulation: Cardiovascular Quality and Outcomes
- Vol. 17 (12) , e010602
Prognostic Significance and Associations of Neural Network–Derived Electrocardiographic Features
BACKGROUND: Subtle, prognostically important ECG features may not be apparent to physicians. In the course of supervised machine learning, thousands of ECG features are identified. These are not limited to conventional ECG parameters and morphology. We aimed to investigate whether neural network–derived ECG features could be used to predict future cardiovascular disease and mortality and have phenotypic and genotypic associations. METHODS: We extracted 5120 neural network–derived ECG features from an artificial intelligence–enabled ECG model trained for 6 simple diagnoses and applied unsupervised machine learning to identify 3 phenogroups. Using the identified phenogroups, we externally validated our findings in 5 diverse cohorts from the United States, Brazil, and the United Kingdom. Data were collected between 2000 and 2023. RESULTS: In total, 1 808 584 patients were included in this study. In the derivation cohort, the 3 phenogroups had significantly different mortality profiles. After adjusting for known covariates, phenogroup B had a 20% increase in long-term mortality compared with phenogroup A (hazard ratio, 1.20 [95% CI, 1.17–1.23]; P P P P P P SCN10A , SCN5A , and CAV1 have roles in cardiac conduction and arrhythmia. ARHGAP24 does not have a clear cardiac role and may be a novel target. CONCLUSIONS: Neural network–derived ECG features can be used to predict all-cause mortality and future cardiovascular diseases. We have identified biologically plausible and novel phenotypic and genotypic associations that describe mechanisms for the increased risk identified.
- 1 December 2024
- Published by Wolters Kluwer Health in Circulation: Cardiovascular Quality and Outcomes
- Vol. 17 (12) , e010602
Open AccessPrognostic Significance and Associations of Neural Network–Derived Electrocardiographic Features
BACKGROUND: Subtle, prognostically important ECG features may not be apparent to physicians. In the course of supervised machine learning, thousands of ECG features are identified. These are not limited to conventional ECG parameters and morphology. We aimed to investigate whether neural network–derived ECG features could be used to predict future cardiovascular disease and mortality and have phenotypic and genotypic associations. METHODS: We extracted 5120 neural network–derived ECG features from an artificial intelligence–enabled ECG model trained for 6 simple diagnoses and applied unsupervised machine learning to identify 3 phenogroups. Using the identified phenogroups, we externally validated our findings in 5 diverse cohorts from the United States, Brazil, and the United Kingdom. Data were collected between 2000 and 2023. RESULTS: In total, 1 808 584 patients were included in this study. In the derivation cohort, the 3 phenogroups had significantly different mortality profiles. After adjusting for known covariates, phenogroup B had a 20% increase in long-term mortality compared with phenogroup A (hazard ratio, 1.20 [95% CI, 1.17–1.23]; P P P P P P SCN10A , SCN5A , and CAV1 have roles in cardiac conduction and arrhythmia. ARHGAP24 does not have a clear cardiac role and may be a novel target. CONCLUSIONS: Neural network–derived ECG features can be used to predict all-cause mortality and future cardiovascular diseases. We have identified biologically plausible and novel phenotypic and genotypic associations that describe mechanisms for the increased risk identified.
- Open Access
- 23 May 2024
- Published by MDPI AG in Electronics
- Vol. 13 (11) , 2043
Assessing the Effects of Various Gaming Platforms on Players’ Affective States and Workloads through Electroencephalogram
Game platforms have different impacts on player experience in terms of affective states and workloads. By studying these impacts, we can uncover detailed aspects of the gaming experience. Traditionally, understanding player experience has relied on subjective methods, such as self-reported surveys, where players reflect on their experience and effort levels. However, complementing these subjective measures with electroencephalogram (EEG) analysis introduces an objective approach to assessing player experience. In this study, we examined player experiences across PlayStation 5, Nintendo Switch, and Meta Quest 2. Using a mixed-methods approach, we merged subjective user assessments with EEG data to investigate brain activity, affective states, and workload during low- and high-stimulation games. We recruited 30 participants to play two games across three platforms. Our findings reveal that there is a statistically significant difference between these three platforms for seven out of nine experience factors. Also, three platforms have different impacts on play experience and brain activity. Additionally, we utilized a linear model to associate player experience aspects such arousal, frustration, and mental workload with different brain regions using EEG data.
- 23 May 2024
- Published by MDPI AG in Electronics
- Vol. 13 (11) , 2043
Open AccessAssessing the Effects of Various Gaming Platforms on Players’ Affective States and Workloads through Electroencephalogram
Game platforms have different impacts on player experience in terms of affective states and workloads. By studying these impacts, we can uncover detailed aspects of the gaming experience. Traditionally, understanding player experience has relied on subjective methods, such as self-reported surveys, where players reflect on their experience and effort levels. However, complementing these subjective measures with electroencephalogram (EEG) analysis introduces an objective approach to assessing player experience. In this study, we examined player experiences across PlayStation 5, Nintendo Switch, and Meta Quest 2. Using a mixed-methods approach, we merged subjective user assessments with EEG data to investigate brain activity, affective states, and workload during low- and high-stimulation games. We recruited 30 participants to play two games across three platforms. Our findings reveal that there is a statistically significant difference between these three platforms for seven out of nine experience factors. Also, three platforms have different impacts on play experience and brain activity. Additionally, we utilized a linear model to associate player experience aspects such arousal, frustration, and mental workload with different brain regions using EEG data.
- Open Access
- 30 August 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (17) , 9825
Deep Representation of EEG Signals Using Spatio-Spectral Feature Images
Modern deep neural networks (DNNs) have shown promising results in brain studies involving multi-channel electroencephalogram (EEG) signals. The representations produced by the layers of a DNN trained on EEG signals remain, however, poorly understood. In this paper, we propose an approach to interpret deep representations of EEG signals. Our approach produces spatio-spectral feature images (SSFIs) that encode the EEG input patterns that activate the neurons in each layer of a DNN. We evaluate our approach using the PhyAAt dataset of multi-channel EEG signals for auditory attention. First, we train the same convolutional neural network (CNN) architecture on 25 separate sets of EEG signals from 25 subjects and conduct individual model analysis and inter-subject dependency analysis. Then we generate the SSFI input patterns that activate the layers of each trained CNN. The generated SSFI patterns can identify the main brain regions involved in a given auditory task. Our results show that low-level CNN features focus on larger regions and high-level features focus on smaller regions. In addition, our approach allows us to discern patterns in different frequency bands. Further SSFI saliency analysis reveals common brain regions associated with a specific activity for each subject. Our approach to investigate deep representations using SSFI can be used to enhance our understanding of the brain activity and effectively realize transfer learning.
- 30 August 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (17) , 9825
Open AccessDeep Representation of EEG Signals Using Spatio-Spectral Feature Images
Modern deep neural networks (DNNs) have shown promising results in brain studies involving multi-channel electroencephalogram (EEG) signals. The representations produced by the layers of a DNN trained on EEG signals remain, however, poorly understood. In this paper, we propose an approach to interpret deep representations of EEG signals. Our approach produces spatio-spectral feature images (SSFIs) that encode the EEG input patterns that activate the neurons in each layer of a DNN. We evaluate our approach using the PhyAAt dataset of multi-channel EEG signals for auditory attention. First, we train the same convolutional neural network (CNN) architecture on 25 separate sets of EEG signals from 25 subjects and conduct individual model analysis and inter-subject dependency analysis. Then we generate the SSFI input patterns that activate the layers of each trained CNN. The generated SSFI patterns can identify the main brain regions involved in a given auditory task. Our results show that low-level CNN features focus on larger regions and high-level features focus on smaller regions. In addition, our approach allows us to discern patterns in different frequency bands. Further SSFI saliency analysis reveals common brain regions associated with a specific activity for each subject. Our approach to investigate deep representations using SSFI can be used to enhance our understanding of the brain activity and effectively realize transfer learning.
- Open Access
- 3 March 2023
- Published by Elsevier in Knowledge-Based Systems
- Vol. 267
Deception detection in conversations using the proximity of linguistic markers
- 3 March 2023
- Published by Elsevier in Knowledge-Based Systems
- Vol. 267
Open AccessDeception detection in conversations using the proximity of linguistic markers
- 7 September 2021
- Published by Springer Nature
- p. 369-380
Resolving Ambiguity in Hedge Detection by Automatic Generation of Linguistic Rules
- 7 September 2021
- Published by Springer Nature
- p. 369-380
Resolving Ambiguity in Hedge Detection by Automatic Generation of Linguistic Rules
- Open Access
- 20 May 2021
- Published by Academic Conferences and Publishing International Ltd in Electronic Journal of e-Learning
- Vol. 19 (3) , pp159-169
Analysis of Factors Affecting the Auditory Attention of Non-native Speakers in e-Learning Environments
: One of the most striking characteristics of e-Learning audiences is their diversity. Native and non-native learners can be expected among such audiences and therefore, when developing e-Learning courses it is important to consider the impact of the language level on learning. Specifically, non-native learners are expected to have a diminished auditory perception compared to native ones and hence reduced attention capabilities that could result in a poorer performance. In this study, we assess the impact of linguistic and auditory factors on the attention of non-native learners, namely semanticity, sentence length and noise level. An English language platform mimicking real e-Learning environments is used and attention is quantified by measuring the number of English words correctly identified during a listening task. Our results show that changes in each factor affect the attention score significantly. Interestingly, the effects of semanticity are apparent in noiseless environment, but vanish in noisy ones. Results also show that in noiseless environments, a change in the length of semantic sentences from small or medium to long causes a significant drop in the attention score. Our results demonstrate the importance of carefully accounting for linguistic and auditory factors when designing effective e-Learning courses, especially when they target global audiences and learners with different language abilities are expected.
- 20 May 2021
- Published by Academic Conferences and Publishing International Ltd in Electronic Journal of e-Learning
- Vol. 19 (3) , pp159-169
Open AccessAnalysis of Factors Affecting the Auditory Attention of Non-native Speakers in e-Learning Environments
: One of the most striking characteristics of e-Learning audiences is their diversity. Native and non-native learners can be expected among such audiences and therefore, when developing e-Learning courses it is important to consider the impact of the language level on learning. Specifically, non-native learners are expected to have a diminished auditory perception compared to native ones and hence reduced attention capabilities that could result in a poorer performance. In this study, we assess the impact of linguistic and auditory factors on the attention of non-native learners, namely semanticity, sentence length and noise level. An English language platform mimicking real e-Learning environments is used and attention is quantified by measuring the number of English words correctly identified during a listening task. Our results show that changes in each factor affect the attention score significantly. Interestingly, the effects of semanticity are apparent in noiseless environment, but vanish in noisy ones. Results also show that in noiseless environments, a change in the length of semantic sentences from small or medium to long causes a significant drop in the attention score. Our results demonstrate the importance of carefully accounting for linguistic and auditory factors when designing effective e-Learning courses, especially when they target global audiences and learners with different language abilities are expected.
- Open Access
- 24 February 2021
- Published by IntechOpen
Wavelets for EEG Analysis
This chapter introduces the applications of wavelet for Electroencephalogram (EEG) signal analysis. First, the overview of EEG signal is discussed to the recording of raw EEG and widely used frequency bands in EEG studies. The chapter then progresses to discuss the common artefacts that contaminate EEG signal while recording. With a short overview of wavelet analysis techniques, namely; Continues Wavelet Transform (CWT), Discrete Wavelet Transform (DWT), and Wavelet Packet Decomposition (WPD), the chapter demonstrates the richness of CWT over conventional time-frequency analysis technique e.g. Short-Time Fourier Transform. Lastly, artefact removal algorithms based on Independent Component Analysis (ICA) and wavelet are discussed and a comparative analysis is demonstrated. The techniques covered in this chapter show that wavelet analysis is well-suited for EEG signals for describing time-localised event. Due to similar nature, wavelet analysis is also suitable for other biomedical signals such as Electrocardiogram and Electromyogram.
- 24 February 2021
- Published by IntechOpen
Open AccessWavelets for EEG Analysis
This chapter introduces the applications of wavelet for Electroencephalogram (EEG) signal analysis. First, the overview of EEG signal is discussed to the recording of raw EEG and widely used frequency bands in EEG studies. The chapter then progresses to discuss the common artefacts that contaminate EEG signal while recording. With a short overview of wavelet analysis techniques, namely; Continues Wavelet Transform (CWT), Discrete Wavelet Transform (DWT), and Wavelet Packet Decomposition (WPD), the chapter demonstrates the richness of CWT over conventional time-frequency analysis technique e.g. Short-Time Fourier Transform. Lastly, artefact removal algorithms based on Independent Component Analysis (ICA) and wavelet are discussed and a comparative analysis is demonstrated. The techniques covered in this chapter show that wavelet analysis is well-suited for EEG signals for describing time-localised event. Due to similar nature, wavelet analysis is also suitable for other biomedical signals such as Electrocardiogram and Electromyogram.