# Rapid invisible frequency tagging (RIFT) does not evoke intermodulation components in the neural response

## TL;DR

This study investigates whether a technique called RIFT can detect nonlinear neural responses during visual integration, but finds it does not produce expected intermodulation components.

## Contribution

The study demonstrates that RIFT does not evoke intermodulation components in neural responses, unlike low-frequency visible tagging methods.

## Key findings

- Participants responded more accurately and faster to coherent motion stimuli compared to incoherent motion.
- Pupil dilation was significantly larger during incoherent motion.
- RIFT did not produce intermodulation components in the neural response, suggesting limitations in its use for studying visual integration.

## Abstract

The human visual system performs nonlinear integrative operations at multiple stages of visual information processing. For instance, integrating parts of visual stimuli into a coherent object involves coordinated neural processing along the visual hierarchy. However, it remains uncertain whether visual integration manifests in a nonlinear neural response, particularly through intermodulation components in the power spectrum. In this study, we used a visual motion paradigm combined with rapid invisible frequency tagging (RIFT) and magnetoencephalography (MEG) to explore nonlinear characteristics of neural responses associated with visual integration. In this paradigm, two grating patches were moving coherently or incoherently, and were modulated by RIFT at 56 and 63 Hz, respectively. The behavioural results revealed that the participants responded more accurately and faster to probes during coherent compared to incoherent motion. Moreover, the type of motion elicited differential effects on pupil dilation, with significantly larger pupil diameter observed during incoherent motion. To evaluate the neural response to coherent and incoherent motion stimuli, we assessed spectral coherence between MEG and RIFT. We observed a strong coherence at the tagging frequencies (f1 = 56 and f2 = 63 Hz) as well as at the higher harmonics at 112 Hz and 126 Hz, respectively. Importantly we did not observe a response at frequencies of the intermodulation (f2–f1, f2 + f1); nor did we observe a difference when comparing the coherence and incoherent motion. We conclude that in contrast to studies with low-frequency visible tagging, RIFT does not evoke intermodulation components and therefore, its applicability for investigating the neural mechanisms of visual integration might be limited.