The dataset has been made publicly available at https://openneuro.org/datasets/ds005700

The purpose of the study is to enable emotion recognition analysis using culturally relevant stimuli, such as Indian Bollywood movie clips, in fMRI data collected from Indian participants. This dataset explores the neural basis of emotions in a contextually meaningful way, focusing on functional and dynamic connectivity.

This dataset contains fMRI data collected from 40 healthy participants who watched emotional Indian movie videos clips. Tasks included resting-state and emotion-elicitation tasks. Data was organized following the BIDS standard. Dataset size: 7.21 GB (BIDS-compliant format), 20 GB (raw DICOM files). The data was collected at fMRI center of Central Institute of Psychiatry (CIP), Kanke, Ranchi.The research proposal and experimental protocol received ethical clearance from the Institute Ethics Committee, CIP (No. IEC/CIP/2022-23/1709).

This dataset was collected by:

Abgeena, Research Scholar, Birla Institute of Technology, Mesra, Ranchi, India

Under the guidance of:

Dr. Shruti Garg, Assistant Professor, Department of Computer Science and Engineering, Birla Institute of Technology, Mesra, Ranchi, India

Dr. Nishant Goyal, Professor, Department of Psychiatry, Central Institute of Psychiatry, Kanke, Ranchi, India

fMRI experimental setup was designed by:

Justin Raj, JRF, Department of Psychiatry, Central Institute of Psychiatry, Kanke, Ranchi, India

Scanner: Philips Ingenia 3T MRI scanner with following parameter- -T1w data was collected for matrix size 192x192x256, voxel size 1x1x1 mm, slice thickness 1, space between slices 1, echo time 0.002943, repetition time 0.0065, flip angle 9°. -Task-rest_bold was collected for matrix size 96x96x38, voxel size -2.29x2.29x4 mm, slice thickness 4, space between slices 4, echo time 0.035001, repetition time 2.02697, flip angle 90°, slice 38. -Task-fe_bold was collected for matrix size 128x128x36, voxel size =-1.8x1.8x4 mm, slice thickness 4, space between slices 4, echo time 0.035, repetition time 3, flip angle 90°, slice 36.

Videos used as stimuli were selected from [Mishra, S., Srinivasan, N., Asif, M., and Tiwary, U.S., 2023. Affective film dataset from India (AFDI): Creation and validation with an Indian sample. Journal of Cultural Cognitive Science, 7(3), pp. 255–267] based on their potential to evoke emotions. The videos were downloaded in high quality and cropped to 30-second clips to elicit specific emotions corresponding to five emotion classes: Calm, Afraid, Delighted, Depressed, and Excited.

Participants viewed movie clips designed to evoke different emotions task details: onset duration emotion

0 30 calm 30 30 white noise 60 30 afraid 90 30 white noise 120 30 delighted 150 30 white noise 180 30 depressed 210 30 white noise 240 30 excited 270 30 white noise 300 30 delighted 330 30 white noise 360 30 depressed 390 30 white noise 420 30 calm 450 30 white noise 480 30 excited 510 30 white noise 540 30 afraid 570 30 white noise

-/sub-01/
-/anat/
-sub-01_T1w.nii: Anatomical data of sub-01
-sub-01_T1w.json
-/func/
    -sub-01_task-fe_bold.nii: Functional data for the emotion task.
-sub-01_task-fe_bold.json
    -sub-01_task-rest_bold.nii: Resting-state functional data.
-sub-01_task-rest_bold.json

For questions regarding dataset feel free to contact at email: [abgeenakhan786@gmail.com]

I extend gratitude to all 40 participants for their valuable time, patience and goodwill in contributing to this data collection.

Please add the following references to your publications related to emotion recognition, if you feel this work is helpful for you.

o Abgeena, A., Garg, S., Goyal, N. and PC, J.R., 2025. NeuroEmo: A neuroimaging-based fMRI dataset to extract temporal affective brain dynamics for Indian movie video clips stimuli using dynamic functional connectivity approach with graph convolution neural network (DFC-GCNN). Computers in Biology and Medicine, 194, p.110439. https://doi.org/10.1016/j.compbiomed.2025.110439

o Abgeena, A. and Garg, S., 2023. A novel convolution bi-directional gated recurrent unit neural network for emotion recognition in multichannel electroencephalogram signals. Technology and Health Care, 31(4), pp.1215-1234. https://doi.org/10.3233/THC-220458

o Abgeena, A. and Garg, S., 2023. S-LSTM-ATT: a hybrid deep learning approach with optimized features for emotion recognition in electroencephalogram. Health Information Science and Systems, 11(1), p.40. https://doi.org/10.1007/s13755-023-00242-x

o Abgeena, A. and Garg, S., 2023, February. EEG evoked automated emotion recognition using deep convolutional neural network. In 2023 Fifth International Conference on Electrical, Computer and Communication Technologies (ICECCT) (pp. 1-7). IEEE. https://doi.org/10.1109/ICECCT56650.2023.10179711