henhaobio
Sign in

bioRxivNeurosciencePreprint

A transient signal in foveal superior colliculus neurons when jumpstarting peripheral saccadic orienting

Tong ZhangiD, Anna F. DenningeriD, Ziad M. HafediD

PDFDOI
In-depth full-text digest available · Jump to digest ↓

Abstract

The superior colliculus (SC) both senses the environment and orients gaze within it. While the SCs sensory and motor bursts appear qualitatively similar to each other, population activity structure in the two processing regimes (visual and motor) is very different, necessitating a hitherto unexplored rapid representational transformation, occurring on the scale of only tens of milliseconds. Here, we first show that when a planned saccade is released with a go signal, peripheral SC neurons representing the saccade target location exhibit a transient, short-latency pause right before their motor bursts eventually emerge. This pause starts within [~]50 ms from the go signal; it is stimulus-dependent; and it is significantly weaker in purely visual SC neurons than in saccade-related ones. Foveal SC neurons, on the other hand, transiently burst, and their bursts lead the peripheral neurons pauses by [~]10 milliseconds. These transient foveal bursts also lead the same foveal neurons saccade-related pauses, which are time-locked to movement generation. Remarkably, during immediate visually-guided saccade tasks, requiring a transformation from visual to motor peripheral bursts in <50-100 ms, the transient foveal SC bursts still occur, resulting in simultaneous short-latency bursting at two disparate SC loci: one foveal; and one eccentric and responding to the visual appearance of the saccade target. Our results suggest that in classic saccade tasks used to investigate visual, motor, and cognitive processes in primate brains, a transient foveal SC signal precedes peripheral saccadic orienting, and may facilitate a necessary rapid representational transformation needed for SC saccade motor bursts to ensue. SignificanceSaccade control studies often involve behavioral paradigms dissociating stimulus presentation, deliberation epochs, and "go" instructions. While the stimulus presentation, deliberation, and actual movement aspects have been well studied, the "go" phases of trials are less considered. We show that foveal superior colliculus neurons exhibit transient activity bursts at the "go" instruction for saccades, likely enabling a rapid representational transformations from a visual to a motor regime in the extrafoveal collicular neurons driving the eye movements.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (12 Jul 2026) · CC BY 4.0

During delayed saccade tasks, macaque foveal superior colliculus neurons burst after the go signal, leading peripheral collicular pauses by 10 ms, independent of target image, saccade direction or reaction time.

Superior colliculus (SC) neurons must switch from visual to motor representations within short intervals in delayed saccade tasks, but earlier work examined peripheral neurons at target onset and saccade initiation rather than at the go instruction itself. By examining foveal as well as peripheral SC at the go signal, this work finds a foveal burst stronger than the subtle, uncharacterized elevations that, per the authors, earlier rostral SC studies only hinted at.

Key results

  1. ShownIn the delayed saccade task, foveal SC neurons fired a transient burst after the go signal (fixation spot removal), the opposite of what peripheral neurons did, and the burst began 10 ms before the peripheral SC pause.

    Normalized foveal population activity rose about three-fold after the go signal while peripheral neurons paused; onsets from pooled spike times in 2-ms bins (first of 3 successive rises or falls) put the burst 10 ms ahead. Burst strength 50-100 ms post-go did not differ between 1 and 4 cpd (cycles/deg) targets (signrank p=0.8509, n=33) and showed no systematic dependence on saccade direction. · Foveal superior colliculus neurons exhibit…; Foveal superior colliculus activity bursts lead…; Saccades-to-gratings task for foveal SC neurons; Figure legends

  2. ShownPeripheral SC neurons paused briefly after the go signal, just before their saccade-related motor bursts. The drop scaled with delay-period activity, so rates converged. It was stronger in visual-motor and motor neurons than in purely visual ones, and it was absent in primary visual cortex (V1).

    Rate 50-150 ms post-go vs final 100 ms pre-go fell in SC (signrank p=7.7693 × 10^-23, 1.4893 × 10^-19, 6.4765 × 10^-18; spatial frequency, contrast, orientation tasks) but not V1 (p=0.6165, 0.4209, 0.5876). Drop larger for most vs least preferred feature (p=3.9233 × 10^-29). Spatial frequency task median modulation index -0.1069 (motor-related) vs -0.0453 (visual), p=0.000532; orientation task p=0.0966. · Peripheral superior colliculus neurons exhibit…; Peripheral superior colliculus activity pauses…; Peripheral superior colliculus activity pauses…; Figure legends

  3. ShownFoveal bursts persisted in an immediate visually-guided saccade task, where fixation spot removal coincided with target onset, so foveal and peripheral SC sites burst at the same time. The authors argue that lateral inhibition alone cannot account for this.

    In population averages, peripheral neurons gave a visual burst and then a saccade motor burst, while foveal neurons burst at the time of the peripheral visual response and then dipped at saccade onset. Foveal neurons lacked saccade-related bursts and their receptive fields (RFs) were checked, arguing against recording inner edges of peripheral RFs. Targets were white discs, not gratings. · Foveal bursts cooccur with peripheral visual…; Immediate visually-guided saccade task for…; Figure legends

  4. SuggestedFoveal go-signal bursts appeared to depend on the instructed-saccade context rather than being simple visual offset responses to fixation spot removal.

    Example neurons showed little RF-mapping offset response at the fixation spot's distance; foveal neurons without go bursts still had offset responses; in bursting neurons burst strength correlated moderately with offset response (r = 0.6026, 0.5924). Removing the foveated grating at trial end (no saccade for 150 ms) evoked no burst, unlike fixation spot removal (signrank p = 6.6424 × 10^-8, 1.5453 × 10^-7). · Foveal superior activity bursts reflect…; Data analysis and statistical tests; Figure legends

  5. SuggestedUnlike peripheral visual responses, foveal bursts did not vary with saccadic reaction time in either task. This fits a go signal that does not depend on saccade timing.

    Trials were split at each session's median reaction time. In the immediate task, peripheral visual bursts were stronger on fast trials (signrank p = 2.6741 × 10^-35, n=253 neurons) but foveal bursts were not (p=0.059, n=41 neurons). Foveal bursts in the delayed task were also similar on fast and slow trials. · Peripheral visual responses, but not foveal…; Figure legends

Methods, data and reproducibility

Re-analysis of spike-sorted SC (peripheral and foveal) and V1 neurons recorded earlier in male rhesus macaques performing delayed saccades to gratings and immediate visually-guided saccades. Go-aligned firing was compared with the 100 ms pre-go baseline, onset timing was taken from 2-ms spike-time histograms, and foveal bursts were compared with offset responses from receptive field mapping.

  • Peripheral SC neurons in the delayed saccades-to-gratings task (spatial frequency, contrast and orientation versions), reused from an earlier study · rhesus macaque (male)
  • Primary visual cortex (V1) neurons in the same delayed saccades-to-gratings task · rhesus macaque (male)
  • Foveal SC neurons in a delayed saccade task with 1 or 4 cpd vertical gratings at about 8 deg eccentricity (control conditions of an earlier study) · rhesus macaque (male) · n = 33 neurons
  • Peripheral SC neurons in an immediate visually-guided saccade task (white disc target) · rhesus macaque (male) · n = 253 neurons
  • Foveal SC neurons in the immediate visually-guided saccade task · rhesus macaque (male) · n = 41 neurons
  • Receptive field mapping trials (white spot at pseudorandom locations) from the same sessions, used for RF maps and stimulus-offset responses · rhesus macaque (male)
Methods & tools
delayed visually-guided saccade paradigm (saccades-to-gratings), immediate visually-guided saccade paradigm, receptive field mapping task, spike sorting (performed in earlier studies), Wilcoxon signed-rank test, Wilcoxon rank sum test, t-test
Reproducibility gaps
No code or data availability statement. The data were collected for earlier publications whose references cannot be resolved in the text. The number of monkeys is not stated, and the neuron counts for the delayed-task peripheral SC and V1 populations are indicated only in the figures.
Limitations (4)
  • Fixation spot removal is itself a foveal visual transient. The authors say it still needs testing whether foveal bursts occur without one, for example by keeping the fixation spot on and cueing the saccade with a spatially uninformative sound. · stated by the authors
  • What drives the foveal bursts, and how they link to peripheral pauses, is left open; the authors propose blocking cortical inputs to the SC to probe this. · stated by the authors
  • No perturbation tests whether foveal bursts trigger peripheral pauses. The 10 ms lead comes from a single pooled spike-time histogram with no per-neuron latency statistic. · noted in this summary
  • An offset-response contribution is not fully excluded. Among bursting foveal neurons, burst strength correlated with RF-mapping offset responses (r = 0.6026 and 0.5924), and one saccade-direction comparison (high spatial frequency) was significant. · noted in this summary

Read this if you record from the superior colliculus or other oculomotor areas in delayed or immediate saccade tasks and want evidence that foveal neurons respond to the go instruction.

Citations
1

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISGopal '25Goffart '17Trottenberg '26Massot '19Khademi '20Zhang '25
79% semantic matchNeuroscience
Atul Gopal, Okihide Hikosaka·iScience·Dec 29, 2025· 1 citations