Temporal dynamics of normalization reweighting

Abstract

For decades, neural suppression in early visual cortex has been thought to be fixed. But recent work has challenged this assumption by showing that suppression can be reweighted based on recent history; when pairs of stimuli are repeatedly presented together, suppression between them strengthens. Here we investigate the temporal dynamics of this process using a steady-state visual evoked potential (SSVEP) paradigm that provides a time-resolved, direct index of suppression between pairs of stimuli flickering at different frequencies (5 and 7 Hz). Our initial analysis of an existing electroencephalography (EEG) dataset (N = 100) indicated that suppression increases substantially during the first 2-5 seconds of stimulus presentation (with some variation across stimulation frequency). We then collected new EEG data (N = 100) replicating this finding for both monocular and dichoptic mask arrangements in a preregistered study designed to measure reweighting. A third experiment (N = 20) used source-localized magnetoencephalography and found that these effects are apparent in primary visual cortex (V1), consistent with results from neurophysiological work. Because long-standing theories propose inhibition/excitation differences in autism, we also compared reweighting between individuals with high versus low autistic traits, and with and without an autism diagnosis, across our three datasets (total N = 220). We find no compelling differences in reweighting that are associated with autism. Our results support the normalization reweighting model and indicate that for prolonged stimulation, increases in suppression occur on the order of 2-5 seconds after stimulus onset.

Figure 1.

Summary of pilot analysis of data from Vilidaite et al. (2018). Panel (a) shows contrast response functions at the target frequency (7 Hz, left) and the mask frequency (5 Hz, right). Insets show the distribution of activity across the scalp, with points marking the electrodes over which signals were averaged (Oz, POz, O1, and O2). Panel (b) shows Fourier spectra for the single-component stimuli and their combination (plaid). Note the strong second harmonic components at 14 Hz and 10 Hz. Panel (c) shows time courses of frequency-locked responses to a single stimulus (black) and the plaid stimulus (blue), compared to a baseline condition (gray) where no stimulus was shown at the target frequency. Panel (d) shows the time course of suppression at each frequency (7 Hz, 5 Hz, 14 Hz, 10 Hz) and their average (black curve). Points around y = 1.8 indicate a significantly increasing ratio (for the time window centered at each point). Error bars in panel (a) and shaded regions in panels (c, d) indicate ±1 SE across N = 100 participants, and gray rectangles indicate the timing of stimulus presentation. The larger symbols in panel (a) indicate the conditions used for subsequent analyses.

Figure 2.

Summary of EEG results for N = 100 adult participants. Panel (a) shows scalp topographies and averaged waveforms for 5 Hz (top) and 7 Hz (bottom) stimuli. The black sine wave trace in each panel illustrates the driving contrast modulation, and black points on the scalp topographies indicate electrodes Oz, O1, O2, and POz. Panel (b) shows the Fourier amplitude spectrum for each condition, with clear peaks at 5 Hz and 7 Hz. Panels (c, d) show time courses at each frequency for the target-only condition (black) and the monocular (blue) and dichoptic (red) masking conditions. Panels (e, f) show suppression ratios as a function of time for each mask type, with points around y = 0.8 indicating a significantly increasing ratio. Shaded regions in panels (c–f) span ±1 SE across participants, and light gray rectangles indicate the period of stimulus presentation.

Figure 3.

Summary of MEG results for N = 20 adults. Panel (a) shows average SSVEP response in source space, thresholded at a signal-to-noise ratio of 2 (blue, left), and the locations of the V1 region of interest on the medial surface of both hemispheres (right, red). Panel (b) shows the Fourier spectra for the four experimental conditions, from the most responsive vertex in V1. Panels (c, d) show time courses at 5 Hz and 7 Hz, and panels (e, f) show suppression ratios for the monocular and dichoptic conditions at each frequency, with points around y = 0.8 indicating a significantly increasing ratio. Shaded regions in panels (c–f) indicate ±1 SE across participants, and light gray rectangles indicate the period of stimulus presentation.

Figure 4.

Time course of the sum intermodulation term at 12 Hz across three experiments. In both EEG experiments, the intermodulation response increases during the first 1 s of stimulus presentation. Responses are calculated as a proportional increase relative to the target-only conditions (where the intermodulation response is absent), for direct comparison with the suppression ratios in Figures 1–3. Shaded regions indicate ±1 SE.

Figure 5.

Analysis of the effect of autistic traits on normalization reweighting. Panels (a–c) show distributions of AQ scores across the three datasets. Panels (d–h) show time courses of suppression averaged across stimulation frequency and split by AQ score (d–f) or autism status (g, h). Panels (d, e, g) are for binocular or monocular presentation, and panels (f, h) are for dichoptic presentation. Shaded regions in panels (d–h) indicate ±1 SE across participants, and black points at y = 0.8 indicate significant differences between groups.