Causal evidence for the role of the sensory visual cortex in visual short-term memory maintenance

Abstract

The role of the sensory visual cortex during visual short-term memory (VSTM) remains controversial. This controversy is possibly due to methodological issues in previous attempts to investigate the effects of transcranial magnetic stimulation (TMS) on VSTM. The aim of this study was to use TMS, while covering previous methodological deficits. Sixty-four young adults were recruited to participate in two experiments (Experiment 1: n = 36; Experiment 2: n = 28) using a VSTM orientation change-detection task under TMS. Monocular vision was ensured using red-blue goggles combined with red-blue stimuli. Double-pulse TMS was delivered at different times (Experiment 1: 0, 200 or 1000 ms; Experiment 2: 200, 1000 ms) during a 2 s maintenance phase, on one side of the occipital hemisphere. In Experiment 2, a sham TMS condition was introduced. Decreased detection sensitivity ( d′ ) in the ipsilateral occipital hemisphere to visual hemifield, and in the real TMS (compared with sham TMS) condition indicated inhibitory TMS effects, and thus, a causal involvement of the sensory visual cortex during early (200 ms) and late (1000 ms) maintenance in VSTM. These findings are aligned with sensory recruitment, which proposes that both perceptual and memory processes rely upon the same neural substrates in the sensory visual cortex. The methods used in this study were preregistered and had received in-principle acceptance on 6 June 2022 (Stage 1 protocol can be found in: https://doi.org/10.17605/OSF.IO/EMPDT ).

The causal involvement of the visual cortex in visual working memory remains uncertain

“…The differences between ipsilateral versus contralateral conditions (in both TMS and sham conditions of experiments 1 and 2, corresponding to H1-H4 and H6; shown in figure 2a,b, respectively) can hence not be attributed to a disruption of VWM in early visual cortex via TMS, but instead to other (unintended) factors differentially affecting the processing of monocularly viewed stimuli that occur during both TMS and sham TMS equally. Interestingly, this null finding regarding the key hypotheses remained largely unacknowledged in the original publication by Phylactou et al [14], potentially due to their rigorous adherence to the analysis protocol that did not plan direct comparisons between TMS and sham in the pre-registered tests. This in turn emphasizes the importance of exploratory analyses.…”

“…Shown is a red stimulus presented at fixation and viewed through red-blue dichoptic goggles, hence selectively reaching the left eye (monocular stimulation). Phylactou et al [14] wrongly assumed that a monocularly presented stimulus at fixation would initially reach the ipsilateral hemisphere (as depicted in the left panel). Instead, the left retina projects signals (via the optic nerve, tract and LGN) bilaterally to the left and right visual cortices.…”

“…Accordingly, any monocular visual stimulus at fixation is actually represented in largely similar patches of early visual areas in both hemispheres (provided participants fixated correctly). (c) Under the mistaken assumption that a monocular stimulus presentation would be processed ipsilaterally, Phylactou et al [14] presented coloured stimuli at fixation that participants viewed with red-blue goggles. As they expected to thereby separate the (initial) processing of the visual stimuli into hemispheres (shown in the left panel), they compared behavioural responses during ipsilateral (experimental condition) and contralateral (control condition) TMS stimulation.…”

“…Phylactou and colleagues wrongly assumed that stimuli presented within 15° of the visual angle (as used in the prior TMS experiments, [5][6][7][8]) would be processed by both hemispheres if viewed with both eyes. For example, in their TMS study, they write: '[…] considering the neuroanatomy of the visual pathway system, the binocular presentation of stimuli either left or right close to the midline of the visual field-as was the case in the majority of the previous studies-does not accurately correspond to the contralateral sensory visual cortex, and could in fact be processed by the ipsilateral cortex if presented within 15° of visual angle from midline [23,24]' ( [14], p. 3). Accordingly, they suggest presenting stimuli monocularly.…”

“…What do these lateralized TMS effects show, and how can we interpret them? To address these questions, we performed a series of reanalyses using the publicly available data from the TMS experiments [14]. We replicated their registered analyses (H1-H7), presented reanalyses including excluded participants and, importantly, performed additional analyses on the data from the second experiment that took into account their control sham condition.…”