Prefrontal fNIRS dataset: functional connectivity during haptic-enabled virtual assembly training at three levels of task complexity

Prefrontal fNIRS dataset: functional connectivity during haptic-enabled virtual assembly training at three levels of task complexity OVERVIEWRaw and Homer3-processed functional near-infrared spectroscopy (fNIRS) recordings from 30 novice participants who performed industrial assembly tasks of graded complexity in a haptic-enabled virtual reality system. The deposit supports the analyses reported in the associated manuscript on the modulation of prefrontal functional connectivity by assembly task complexity. PARTICIPANTS AND DESIGNWithin-subject design with three conditions: resting state (7 min, eyes closed), low-complexity assembly (oil pump, 5 parts) and high-complexity assembly (linear actuator, 8 parts), each with a 30 min ceiling. Thirty final-semester mechanical engineering students (21 men, 9 women; mean age 21.93 years, s = 1.79), all without prior experience of haptic-enabled VR. Total session duration 92 min, without breaks. Task order was not counterbalanced. ACQUISITIONPortable Brite MKII system (Artinis Medical Systems B.V., Elst, the Netherlands); ten dual-wavelength sources (757 and 843 nm) and eight detectors; sampling rate 25 Hz. The montage yields 22 long channels (3 cm source-detector separation) covering the prefrontal cortex and 2 short channels (1.5 cm) for superficial signal regression. Virtual assembly was executed in the Haptic Assembly and Manufacturing System (HAMS) with a Phantom Omni six-degree-of-freedom haptic device. CONTENTS AND ORGANISATIONThe archive is organised as three self-contained, condition-level datasets, each following a BIDS-compatible layout: resting_state/ low_complexity_task/ (oil pump, 5 parts) high_complexity_task/ (linear actuator, 8 parts) Every condition folder has the same internal structure: <condition>/ bids_compatible/ sub-01/ ... sub-30/ nirs/ raw recordings (SNIRF) derivatives/ homer/ sub-01/ ... sub-30/ Homer3 v1.54.0 processed output Thirty participants are present in all three conditions, giving 90 recording sessions. Because each condition is self-contained, a single condition may be downloaded and analysed independently. Group-level connectivity products (rMatFDR, zMatFDR, pMatFDR, pMat, keepRun for HbO, HbR and HbT) are supplied as MATLAB .mat files at the root of the archive.Video recordings of each subject for all 3 conditions are available at: https://www.youtube.com/playlist?list=PLVLvlpQHruzw CORRESPONDENCE WITH THE ANALYSIS CODEThe scripts in the companion repository refer to the three conditions as resting, easy and hard. The mapping is: resting -> resting_state (resting state, 7 min, eyes closed) easy -> low_complexity_task (low-complexity assembly) hard -> high_complexity_task (high-complexity assembly) Set the input directories in the first script to <condition>/bids_compatible/derivatives/homer. PROCESSING SUMMARYChannel pruning (hmrR_PruneChannels; range 1e-7 to 1e7, SNRthresh = 2, SDrange = 0.45), conversion to optical density, per-channel motion artefact detection (AMPthresh = 0.5, SDthresh = 15, tMotion = 1 s, tMask = 1 s), spline-Savitzky-Golay motion correction (p = 0.99, FrameSize_Sec = 10 s), band-pass filtering in the functional connectivity band (0.009 to 0.08 Hz), and conversion to haemoglobin concentration through the modified Beer-Lambert law (partial pathlength factor of 1 per wavelength). Short-channel regression used a Tikhonov-regularised estimator. Connectivity was quantified as Pearson correlation between all channel pairs, Fisher z-transformed, with false discovery rate control. Group inference used repeated-measures ANOVA on subject-level mean z-scores with Tukey-Kramer post hoc comparisons (alpha = 0.05). ANALYSIS CODEThe MATLAB scripts that reproduce the connectivity matrices, statistical contrasts, hemispheric comparisons and figures are archived in the companion repository (see Related identifiers). External dependencies: Homer3 and the Network Based Statistic toolbox. ETHICSThe study followed the Declaration of Helsinki and institutional data-confidentiality policies. Electronic informed consent was obtained before participation and data were handled anonymously. FUNDINGSupported by two postdoctoral scholarships from SECIHTI (Secretariat of Science, Humanities, Technology and Innovation of Mexico). LICENCECreative Commons Attribution 4.0 International (CC BY 4.0).