A Lightweight Transformer for Pain Recognition from Brain Activity

TL;DR

This paper introduces a lightweight transformer model that effectively fuses multiple fNIRS signal representations for real-time pain recognition, balancing accuracy and computational efficiency.

Contribution

A novel lightweight transformer architecture that jointly models heterogeneous fNIRS signals without increasing complexity or requiring modality-specific adjustments.

Findings

Achieves competitive pain recognition accuracy on AI4Pain dataset.

Maintains computational efficiency suitable for real-time deployment.

Effectively fuses multiple signal representations through a unified tokenization mechanism.

Abstract

Pain is a multifaceted and widespread phenomenon with substantial clinical and societal burden, making reliable automated assessment a critical objective. This paper presents a lightweight transformer architecture that fuses multiple fNIRS representations through a unified tokenization mechanism, enabling joint modeling of complementary signal views without requiring modality-specific adaptations or increasing architectural complexity. The proposed token-mixing strategy preserves spatial, temporal, and time-frequency characteristics by projecting heterogeneous inputs onto a shared latent representation, using a structured segmentation scheme to control the granularity of local aggregation and global interaction. The model is evaluated on the AI4Pain dataset using stacked raw waveform and power spectral density representations of fNIRS inputs. Experimental results demonstrate competitive pain recognition performance while remaining computationally compact, making the approach suitable for real-time inference on both GPU and CPU hardware.