A Novel Framework for Unconscious Processing

Abstract

Understanding the distinction between conscious and unconscious cognition remains a priority in psychology and neuroscience. A comprehensive neurocognitive account of conscious awareness will not be possible without a sound framework to isolate and understand unconscious information processing. Here we provide a brain-based framework that allows the identification of unconscious processes even with null effects on behaviour.

Informative neural representations of unseen contents during higher-order processing in human brains and deep artificial networks

“…We tested a high-precision, within-subject framework to provide a representational account of the scope of information processing for unseen items, even those associated with null perceptual sensitivity (Soto et al, 2019) in both brains and deep artificial neural networks. Isolating the brain representation of unconscious contents has been difficult to achieve in systematic and reliable fashion in previous work, with low numbers of trials and signal detection theoretic constraints (Macmillan, 1986) not allowing to decisively discard conscious perception (Fang & He, 2005; Hesselmann et al, 2011; Gayet et al, 2020; Ludwig et al, 2015), and, critically, when unconscious content could be decoded, this was restricted to visual cortex - see also (Ludwig & Hesselmann, 2015).…”

Feature-Based Attentional Weighting and Re-weighting in the Absence of Visual Awareness

“…This way, one ensures that the stimulus perception and the effect of the stimulus are measured in the same context (e.g., Avneon and Lamy, 2018). Second, studies that define unconscious processing only employing subjective awareness measures (e.g., Cheesman and Merikle, 1986) suffer from the criterion problem that arises when conscious knowledge is held with low confidence, hence objective measures that can ensure a clear absence of visual awareness (i.e., if d' = 0) are critical to studying unconscious information processing, which would then be pinpointed by informationbased analyses of neural measures (Soto et al, 2019). Yet, to come up with an exhaustive means that measures visual awareness and unawareness equally well, the joint use of both the objective and subjective measures seems optimal (e.g., Wiens, 2007).…”

“…Thus, particularly switch trials in which the prior orientation was the highly frequent tilt should show prolonged response latencies on the behavioral level, given that the increased likelihood of one orientation over the others was sufficient to induce a prior selection bias (e.g., Leber et al, 2009;Chetverikov et al, 2017). Importantly, a combination of signal detection theoretic measures (Stanislaw and Todorov, 1999) and subjective perceptual ratings (Ramsøy and Overgaard, 2004) was used to assess participants' awareness of the stimulus to avoid potential confounds due to criterion biases in reporting (un)awareness, e.g., reports of no experience for the knowledge held with low confidence (Wiens, 2007;Soto et al, 2019). Therefore the unconscious reweighting of selection hypothesis was eventually tested by maintaining a clear separation between the measures of selective attention weighting, inferred by the pattern of response latencies, and the measures that we used to probe (un)awareness of the stimulus (objective orientation discrimination task and subjective reports).…”

Similar history biases for distinct prospective decisions of self-performance

“…Hence, the results are in keeping with the proposal that metacognitive confidence and visual awareness can be dissociated (see also 22 ) and with the view that higher-order cognitive processes are to some extent dissociable from conscious experience 11,12,[23][24][25] . However, further studies are needed in which metacognitive processing of perceptual decision making is assessed using both neural and behavioural measures under experimental conditions associated with null perceptual sensitivity [26][27][28] .…”