Distinguishing expectation and attention effects in processing temporal patterns of visual input

Abstract

The current study investigated how the brain sets up expectations from stimulus regularities by evaluating the neural responses to expectations driven implicitly (by the stimuli themselves) and explicitly (by task demands). How the brain uses prior information to create expectations and what role attention plays in forming or holding predictions to efficiently respond to incoming sensory information is still debated. We presented temporal patterns of visual input while recording EEG under two different task conditions. When the patterns were task-relevant and pattern recognition was required to perform the button press task, three different event-related brain potentials (ERPs) were elicited, each reflecting a different aspect of pattern expectation. In contrast, when the patterns were task-irrelevant, none of the neural indicators of pattern recognition or pattern violation detection were observed to the same temporally structured sequences. Thus, results revealed a clear distinction between expectation and attention that was prompted by task requirements. These results provide complementary pieces of evidence that implicit exposure to a stimulus pattern may not be sufficient to drive neural effects of expectations that lead to predictive error responses. Task-driven attentional control can dissociate from stimulus-driven expectations, to effectively minimize distracting information and maximize attentional regulation.

Keywords: Event-related potentials (ERPs); Expectation; Predictive coding; Temporal processing; Visual attention.

Fig. 1.. Stimuli and experimental design.

(A) Standard temporal pattern. Two gratings of different orientations (horizontal, 0° and vertical, 90°) were presented in a temporal sequence: one grating appearing on the screen for 200 ms, followed by a blank screen for 500 ms in a fixed temporal order of three horizontal gratings followed by one vertical grating. This was the standard repeating pattern. (B) Standard pattern and pattern violations. The standard pattern (STD) was presented frequently. The two pattern violations were created by misplacing the temporal position of the vertical grating. The vertical grating occurred after two horizontal gratings (earlier than expected, ED) or after four horizontal gratings (later than expected, LD). The two pattern violations were presented randomly, and each violation type replaced 10 % of the STD patterns. Globally, the horizontal grating appeared 75 % and the vertical grating 25 %.

Fig. 2.. Standard pattern response/Contingency Response (CR).

(A) ERP responses to the repeating STD pattern. Grand-mean waveforms time-locked to the onset of the standard temporal patterns are displayed at the Oz electrode for the conditions: pattern relevant (PR, solid black line) and the pattern irrelevant (PiR, solid gray line). The timing of the stimulus presentation is denoted by the gratings placed above the graph and with the timing in milliseconds displayed along the x-axis. The amplitude is displayed in microvolts on the y-axis. (B) Scalp voltage maps. The scalp voltage maps reflect the difference between the O_ED-O_std and the X₄-X₃ ERP waveforms and are displayed (with a scaling of 0.13 μV per step) at the peak of the interval (mid-point) between the onsets of the successive stimuli, separately for the conditions: PiR (top row) and PR (bottom row). These are labeled in successive order for contingency response across the patterned stimulus presentation relative to the onset of the standard X₁ stimulus (displayed in panel A above) as 1 (600 ms), 2 (1300 ms), 3 (2000 ms), 4 (2700 ms), and 5 (3400 ms), which faithfully reflects the 700 ms SOA. Interval #5 is to show the successive response at the start of the next STD pattern, thus effectively showing the same interval as #1 (see Section 3.2.1 for further details). (C) Mean amplitudes and standard deviations. The grand-mean amplitudes of the waveforms in the 100 ms interval measured between gratings of the standard patterns are displayed along with their standard deviations for the PR (solid black line) and PiR (solid gray line) conditions overlain. Note the return to zero baseline for the #4 interval – the resolution of the STD pattern – prior to the start of the next STD pattern. The negative displacement during the interval between stimulus presentations is clearly seen in the waveforms (A) and in the scalp voltage maps showing the topographic distribution (B).

Fig. 3.. Pattern violation/change detection response.

The grand-mean waveforms (at Oz) time-locked to the O stimulus of the STD pattern (gray, solid line) and the O stimulus of the ED violation (orange, solid line) (ED) are displayed in the left column. Grand-mean waveforms (at Oz) time-locked to the X3 stimulus of the STD pattern (black, solid line) and the X4 stimulus of the LD violation (blue, solid line) are displayed in the right column. Scalp voltage maps (scaled at 0.20 μV per step, negative in blue, positive in red) are displayed at the peak of change response coinciding with the early deviant (O, left panel) and the late deviant (X₄, right panel). (A) Pattern Relevant change responses. A pattern violation response – - a negative displacement of the waveform compared to the standard – - was elicited only during the pattern relevant task (PR, Top row). (B) Pattern irrelevant change responses. No difference between the STD and ED or LD violation responses was observed to the same stimulus sequences when the patterning of the stimuli was irrelevant to the task (PiR, bottom row). Note that the response to the same orientation grating is displayed when it was in the standard pattern and when it was in the deviant pattern, to reveal pattern violation detection. The late deviant is a repetition of the same orientation grating in both conditions. Only in the Pattern Relevant condition is there a negative difference occurring bilaterally at occipital electrodes (compare panels A and B, 330 ms for ED and 300 ms for LD). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4.. P3b target detection response.

(A) Pattern Relevant condition. The grand-mean waveforms time-locked to the onset of the first grating of the STD (black line), ED (orange line), and LD (blue line) patterns for the Pz electrode (top row) and Oz electrode (second row). The task was to press the response key when pattern violations were detected. Significant P3b components were observed at the Pz electrode, with scalp voltage maps (0.25 μV per step, red is positive/blue is negative) displayed at the peak of each P3b component, for the ED (2045 ms) and the LD (2685 ms) responses. Pattern Irrelevant condition. The grand-mean waveforms time-locked to the onset of the first grating of the STD (black line), ED (orange line), and LD (blue line) patterns are displayed at Pz (top row) and Oz (second row) electrodes. The task was to count each grating as it appeared temporally on the computer monitor. P3b was not elicited by the ED or LD pattern violations when the viewer was counting the stimuli. Scalp voltage maps are displayed at the peak location where P3b components were elicited during the pattern task (seen in panel A). The expected scalp distribution with a positive peak focused at the Pz electrode is seen in the Pattern Relevant (panel A) and is not observed for the same stimuli in the Pattern Irrelevant condition. The y-axis displays amplitude in microvolts and the x-axis displays time in milliseconds. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)