A Neural Mechanism of Pain Onset and Offset
Poster No:
1600
Submission Type:
Abstract Submission
Authors:
Jungwoo Kim1, Jiwoong Park1, Christopher Whyte2, Choong-Wan Woo1, James Shine2, Seng Bum Yoo1
Institutions:
1Sungkyunkwan University, Suwon, South Korea, 2The University of Sydney, Sydney, NSW, Australia
First Author:
Co-Author(s):
Introduction:
As a thermal stimulus increases in intensity, our perception shifts from warmth to pain at a certain point. How does the brain translate this quantitative change into a qualitative change in perception? Previous studies have suggested that multivariate brain activity patterns can classify the difference between painful and non-painful perception (Lee et al., 2024), but the computational mechanism underlying how the brain enables this perceptual shift remains unclear. Here, we adopt a dynamical systems perspective to understand this transition with human fMRI. While participants experience temperature-varying thermal stimuli, we examined the changes in the neural dynamics around pain onset and offset. We further trained a recurrent neural network (RNN) model that mimics participants' pain reports and compared its hidden layer dynamics with the neuronal dynamics, uncovering mechanisms underlying a qualitative shift of pain.
Methods:
We delivered thermal stimuli that started at a warmth level, increased to a painful level, and returned to the warmth level across two sessions (n = 24). In Session 1 (mockup scanner), participants continuously rated perceived intensity with a cursor and reported pain onset/offset via button clicks. In Session 2 (3T MRI scanner), they reported only pain onset/offset via button clicks (Fig. 1a). With multi-echo fMRI data, we applied a finite impulse response (FIR) model to estimate the temporal activity of each trial. The averaged brain activity during the constant thermal stimuli (E1, E3, and E5) were tested against two different models: one encoding stimulus intensity and the other encoding warmth vs. painful perception (Fig. 1b). We defined these models as representational dissimilarity matrices (RDMs), which capture pairwise distances between conditions based on hypothesized encoding structure (Nelli et al. 2023). We then identified brain regions where neural RDMs fit these model RDMs using a searchlight analysis (Fig. 2a). Within brain regions that significantly fit both RDM models, we averaged FIR time-series across participants and then concatenated across conditions. We then applied principal component analysis (PCA) to the concatenated data to examine the low-dimensional neuronal dynamics. For the RNN model, the input consisted of the actual temperature levels for the nine experimental conditions with training labels comprising condition-specific intensity ratings and button-click timings from participants' pain reports. During each training epoch, Gaussian noise was added to condition-averaged intensity ratings, and button-click timings were sampled from Gaussian distributions fitted to participants' first and second click timings. The model was trained using backpropagation with mean squared error as the loss function to optimize parameters.
Results:
For the behavioral data, we averaged continuous pain ratings and click timings across trials for the warm and pain conditions (Fig. 1c). Behavioral ratings showed significant differences across the temperature levels (Fig. 1d), whereas click timings showed significant differences across the warm conditions but not across the pain conditions (Fig. 1e).
We then identified brain regions encoding intensity and warmth vs. pain information by fitting the model RDMs onto the neural RDMs (Figs. 2 b-c). The brain regions that showed significant results for both mode RDMs included the anterior cingulate cortex and insula (Fig. 2d). We then visualized the low-dimensional neural dynamics of brain activity of these regions for each condition (Fig. 2e). We also examined the RNN model activations, which displayed patterns similar to the neural representations of stimulus information (Fig. 2f).
We then identified brain regions encoding intensity and warmth vs. pain information by fitting the model RDMs onto the neural RDMs (Figs. 2 b-c). The brain regions that showed significant results for both mode RDMs included the anterior cingulate cortex and insula (Fig. 2d). We then visualized the low-dimensional neural dynamics of brain activity of these regions for each condition (Fig. 2e). We also examined the RNN model activations, which displayed patterns similar to the neural representations of stimulus information (Fig. 2f).
Conclusions:
In this study, we identified brain regions that encode both intensity and pain information. By comparing the neural dynamics of the brain regions with the RNN dynamics, which mimics pain reports, we provide computational insights into the neural mechanisms underlying the feeling of pain.
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI)
Multivariate Approaches 1
Perception, Attention and Motor Behavior:
Perception: Pain and Visceral 2
Keywords:
Computational Neuroscience
Multivariate
Pain
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Nelli, S. (2023). Neural knowledge assembly in humans and neural networks. Neuron, 111(9), 1504-1516.e9.