Related Experiment Video
Updated: Jan 16, 2026
10:36
fMRI Validation of fNIRS Measurements During a Naturalistic Task
Published on: June 15, 2015
21.5K
DeepPhysioRecon: Tracing peripheral physiology in low frequency fMRI dynamics
Roza G Bayrak1,2, Colin B Hansen2,3, Jorge A Salas1,4
1Department of Electrical and Computer Engineering, Vanderbilt University, Nashville, TN, United States.
Imaging Neuroscience (Cambridge, Mass.)
|September 29, 2025
Summary
This study introduces DeepPhysioRecon, a novel AI tool that extracts physiological data like breathing and heart rate from brain imaging (fMRI). This enhances fMRI analysis by revealing crucial brain-body interactions.
Area of Science:
- Neuroimaging and Computational Neuroscience
- Physiological Monitoring and Signal Processing
Background:
- Functional magnetic resonance imaging (fMRI) studies often omit vital physiological measurements.
- Autonomic physiological fluctuations (respiration, heart rate) significantly influence fMRI signals and brain function.
- Lack of physiological data limits the interpretation and depth of fMRI research.
Purpose of the Study:
- To develop a method for decoding physiological variations from fMRI data.
- To assess the generalizability and utility of the developed approach across different datasets and conditions.
- To underscore the importance of integrating physiological measures into fMRI analyses.
Main Methods:
- Developed DeepPhysioRecon, a Long-Short-Term-Memory (LSTM)-based neural network.
- The network decodes continuous respiration amplitude and heart rate from whole-brain fMRI dynamics.
- Conducted systematic evaluations to test generalizability across datasets and experimental settings.
Main Results:
- Successfully decoded continuous physiological variations (respiration, heart rate) from fMRI data.
- Demonstrated the generalizability of the DeepPhysioRecon approach.
- Quantified the impact of including physiological measures in fMRI analyses.
Conclusions:
- DeepPhysioRecon provides a powerful tool for studying brain-body interactions.
- Integrating physiological decoding can enhance fMRI's efficacy as a biomarker.
- The open-source software promotes wider application in neuroscience research.
