Abstract
It has been theorized that cortical feed-forward and recurrent neural activity support unconscious and conscious cognitive processes, respectively. Here we causally tested this proposition by applying event-related transcranial magnetic stimulation (TMS) at early and late times relative to visual stimuli, together with a pulse designed to suppress conscious detection. Consistent with pre-registered hypotheses, early TMS affected residual, reportedly ‘unseen’ capacity. However, conscious perception also appeared critically dependent upon feed-forward processing to a greater extent than the later recurrent phase. Additional exploratory analyses suggested that these early effects dissociated from top-down criterion measures, which were most affected by later TMS. These findings are inconsistent with a simple dichotomy where feed-forward and recurrent processes correspond to unconscious and conscious mechanisms. Instead, different components of awareness may correspond to different phases of cortical dynamics in which initial processing is broadly perceptual whereas later recurrent processing might relate to decision to report.
Highlights
One of the most influential functional descriptions of how and when consciousness manifests in the human brain is the suggestion that the initial feed-forward and later recurrent sweeps of activity support unconscious and conscious processing, respectively (Lamme and Roelfsema 2000; Lamme et al 2000; Lamme 2001, 2006a)
Consistent with our baseline hypothesis, transcranial magnetic stimulation (TMS) applied in isolation at 110 ms after stimulus onset produced evidence of a significant suppression of conscious detection (active vs. sham primary conscious detection sensitivity measure (PrC), T(48) 1⁄4 À3.03, P 1⁄4 0.004, mean 1⁄4 À0.08, 95% CI [À0.13, À0.03], d 1⁄4 0.43, BFuni 1⁄4 12.46, Bayesian test with a default JZS prior (BFjzs) 1⁄4 8.44) while ‘unseen’ discrimination (PcU) remained above chance (with pre-registered exclusions T(40) 1⁄4 6.82, P 1⁄4 3.33 Â 10À8, mean 1⁄4 0.65, 95% CI [0.61, 0.69], d 1⁄4 1.07, BFuni 1⁄4 2.97 Â 109, BFjzs 1⁄4 4.12 Â 105, without exclusions T(49) 1⁄4 6.91, P 1⁄4 9.07 Â 10À9, mean 1⁄4 0.65, 95% CI [0.61, 0.70], d 1⁄4 0.98, BFuni 1⁄4 6.51 Â 109, BFjzs 1⁄4 1.38 Â 106)
Further pre-registered analyses in which individual TMS onset latencies are analysed separately revealed strong evidence that TMS applied at the earliest time (30 ms) suppressed PcU compared to its later TMS counterpart at 190 ms (T(40) 1⁄4 À3.57, P 1⁄4 9.39 Â 10À4, mean 1⁄4 0.08, 95% CI [0.34, 0.12], d 1⁄4 0.56, BFmain 1⁄4 89.37, BFuni 1⁄4 135.95, BFjzs 1⁄4 32.05, see Fig. 2A)
Summary
One of the most influential functional descriptions of how and when consciousness manifests in the human brain is the suggestion that the initial feed-forward and later recurrent sweeps of activity support unconscious and conscious processing, respectively (Lamme and Roelfsema 2000; Lamme et al 2000; Lamme 2001, 2006a). Correlational evidence links early electrophysiological visual components to reportedly ‘unseen’ masked stimuli (Fahrenfort et al 2007). Direct causal evidence, involving a manipulation of ‘unseen’ capacity, is lacking though there have been unsuccessful attempts to manipulate it (Koivisto et al 2011; Koenig and Ro 2019). Disruption of cortical activity very early on, relative to stimuli presentation, has not Received: 11 October 2019; Revised: 30 March 2020.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
