Siyun Liu¹ (), Sheng He; ¹Insititute of Biophysics, Chinese Academy of Sciences

Invisible stimuli have the potential to influence cognitive processes without entering conscious awareness. Previous studies suggested that invisible stimuli can even affect the contents of working memory, but the claim of stimulus invisibility requires more vigorous validation. Here we employed a 2-alternative-forced-choice paradigm to validate invisibility and explored how invisible stimuli affect visual working memory. In a working memory task, participants (n=21, 9 males) briefly viewed a sample figure for 0.5 seconds and memorized it. After a two-second blank screen, participants were presented with noise, visible, or invisible interfering figures for four seconds. Invisibility was achieved through continuous flashing suppression (CFS). The number of interfering figures varied from two to four, randomly. While these interference figures shared the sample's layout, they were never identical. Following another two-second blank screen, a probe figure was presented. The participants evaluated and responded as soon as possible whether the probe matched the sample. The probes were different from the interfering figures in all trials. Three validation blocks were conducted to check the invisibility of the suppressed interfering images, with participants making forced-choice judgments about their location. Accuracy remained at chance levels, confirming invisibility. Participants' normalized reaction times in different conditions showed that invisible stimuli increased reaction times compared to noise, particularly when the probe matched the sample (t(1789) = 2.90, p = .003). The number of stimuli figures did not yield significant differences in reaction times. Our study reveals that unconscious exposure to invisible stimuli sharing the layout as the content in the working memory can delay reaction times in subsequent match-to-sample task. When the invisible interfering pattern differed from the working memory content, participants exhibited slower responses in determining a match between the probe and the sample.