Distinguishing the Neural Correlates of Perceptual Awareness and Postperceptual Processing

Abstract

To identify the neural correlates of perceptual awareness, researchers often compare the differences in neural activation between conditions in which an observer is or is not aware of a stimulus. While intuitive, this approach often contains a critical limitation: to link brain activity with perceptual awareness, observers traditionally report the contents of their perceptual experience. However, relying on observers' reports is problematic because it is difficult to know whether the neural responses being measured are associated with conscious perception or with postperceptual processes involved in the reporting task (e.g., working memory, decision-making). To address this issue, we combined a standard visual masking paradigm with a recently developed "no-report" paradigm in male/female human participants. In the visual masking paradigm, observers saw images of animals and objects that were visible or invisible, depending on their proximity to masks. Meanwhile, on half of the trials, observers reported the contents of their perceptual experience (i.e., report condition), while on the other half of trials they refrained from reporting about their experiences (i.e., no-report condition). We used electroencephalography to examine how visibility interacts with reporting by measuring the P3b event-related potential, one of the proposed canonical "signatures" of conscious processing. Overall, we found a robust P3b in the report condition, but no P3b whatsoever in the no-report condition. This finding suggests that the P3b itself is not a neural signature of conscious processing and highlights the importance of carefully distinguishing the neural correlates of perceptual awareness from postperceptual processing.SIGNIFICANCE STATEMENT What are the neural signatures that differentiate conscious and unconscious processing in the brain? Perhaps the most well established candidate signature is the P3b event-related potential, a late slow wave that appears when observers are aware of a stimulus, but disappears when a stimulus fails to reach awareness. Here, however, we found that the P3b does not track what observers are perceiving, but instead tracks what observers are reporting. When observers are aware of simple visual stimuli, the P3b is nowhere to be found unless observers are reporting the contents of their experience. These results challenge the well established notion of the P3b as a neural marker of awareness and highlight the need for new approaches to the neuroscience of consciousness.

Keywords: EEG; attention; awareness; consciousness; perception; vision.

Figure 1.

Design of experiment 1. Stimuli (i.e., animals or objects) or blank displays were presented in between masks. a, On visible trials, there were 200 ms gaps separating the stimuli from the masks. b, On masked trials, the masks came immediately before and after the stimulus, rendering them completely invisible. c, In the report condition, participants reported on a trial-by-trial basis whether they saw an animal, an object, or nothing. In the no-report condition, the stimulus presentation sequence was the same, but instead of reporting on these stimuli, participants counted the number of times they saw a green circle and reported their count at the end of each block.

Figure 2.

Behavioral results from experiment 1. In all plots, percentage correct (i.e., performance) is plotted on the y-axis. a, Performance on the animal/object/nothing task in the report condition. On the x-axis are the different experimental conditions corresponding to when the target stimulus was visible, invisible (i.e., masked), or absent (i.e., blank). b, Performance on the green circle counting task in the no-report condition. c, Performance on the incidental memory test in the no-report condition for the stimuli that were visible or masked. All error bars represent the standard deviation.

Figure 3.

a, b, ERP results for both the report (top row; a) and the no-report (bottom row; b) conditions. For both condition, topographical voltage distributions over a series of time windows (difference between visible and masked) and the waveforms (for both visible and masked stimuli) from a pool of central-parietal electrodes are plotted. A clear P3b was present in the report condition when observers were aware of the task-relevant stimulus, but the P3b completely vanished in the no-report condition when these same stimuli were task irrelevant. Amplitude scales for the topography maps are as follows: ±4 µV (P1); ±5 µV (N1/P2); ±6 µV (P3b in report condition); ±4 µV (P3b in no-report condition).

Figure 4.

a, b, Results from the mass univariate analyses for both the report (a) and the no-report (b) condition. Each individual electrode is plotted as a row on the y-axis, while time (in milliseconds) is plotted on the x-axis. Only significant t values (5% FDR) are plotted in the figure.

Figure 5.

Design of experiment 2. a, Stimuli (i.e., animals or objects) or blank displays were presented immediately after the fixation dot disappeared. b, In the report condition, participants reported on a trial-by-trial basis whether they saw an animal, object, or nothing. In the no-report condition, the stimulus presentation sequence was the same, but instead of reporting on these stimuli, participants counted the number of times they saw a green circle and reported their count at the end of each block.

Figure 6.

Behavioral results from experiment 2. In all plots, the percentage correct (i.e., performance) is plotted on the y-axis. a, Performance on the animal/object/nothing task in the report condition. On the x-axis are the different experimental conditions corresponding to when the target stimulus was present or absent (i.e., blank). b, Performance on the green circle counting task in the no-report condition. c, Performance on the incidental memory task in the no-report condition for both word and picture stimuli. All error bars represent the standard deviation.

Figure 7.

a, b, ERP results for both the report (top row; a) and the no-report (bottom row; b) conditions. For both conditions, topographical voltage distributions over a series of time windows (difference between visible and masked) and the waveforms (for both visible and masked stimuli) from a representative electrode (Pz) are plotted. A clear P3b was present in the report condition when observers were presented a task-relevant stimulus, but the P3b completely vanished in the no-report condition when these same stimuli were task irrelevant. Amplitude scales for the topography maps are as follows: ±4 µV (P1); ±5 µV (N1/P2); ±4 µV (P3b). Note that in this experiment, a posterior N1 was elicited by both stimuli and blanks (fixation dot offset), thus N1 amplitude differences (stimulus minus blank) were much smaller than in experiment 1.

Figure 8.

a, b, Results from the mass univariate analyses for both the report (a) and the no-report (b) condition. Each individual electrode is plotted as a row on the y-axis, while time (in milliseconds) is plotted on the x-axis. Only significant t values (5% FDR) are plotted in the figure. Note that the absence of a posterior N1 (160–200 ms) was due to both stimulus and blank trials eliciting an N1.

Figure 9.

ERP differences during late time windows revealed by exploratory analyses for both experiment 1 (visible vs masked stimuli; left panels) and experiment 2 (stimuli vs blanks; right panels). For both experiments, topographical differential voltage distributions over a series of time windows starting at 300 ms and ending at 1000 ms are shown for both the report (top: a, masking c, no masking) and the no-report (bottom: b, masking and d, no masking) conditions. Amplitude scales for the topography maps are as follows: 66 mV (300–600 ms in the masking report condition); 64 mV (all other maps).

Figure 10.

Source estimations of the difference between visible and masked stimuli in experiment 1 for both the report (top row) and the no-report (bottom row) conditions across three late time windows (Fig. 9, average voltage distributions during these same time windows). Only significant t values (5% FDR) are plotted on the lateral surface of the Montreal Neurologic Institute brain.