Perception, rather than a continuous combination of sensory input with prior concept, appears to results from sampling of inputs (VanRullen & Koch, 2003) at a frequency in the theta band. Also, binocular rivalry dominance durations can be manipulated using auditory stimulation (Alais et al, 2010). Thus, using techniques of psychophysics and magnetoencephalography (MEG), we tested the notion of an "attentional strobe" and the degree to which it could be manipulated using auditory stimulation. Seven males and eight females undertook two visual change detection tasks. The “rapid” task measured contrast threshold for an abrupt change of one of roughly 15 coloured dots in a standard gap paradigm, while the “gradual” task required the detection in a similar array of dots, of a dot very slowly changing in contrast. Two auditory conditions were employed – 10Hz tone bursts (3.3kHz, duration 5ms, sound level 52dB) or no auditory stimulation. Threshold contrast decreased with 10Hz stimulation for the rapid task while threshold frequency increased for the gradual task with 10Hz stimulation, both consistent with increasing the attentional strobe rate [Task (rapid, gradual) x Auditory(10Hz, no sound) interaction (F(1,120)=7.6, p=.012)]. Theta band power difference of MEG planar gradiometers over posterior parietal sensors of right hemisphere showed opposite effects of auditory stimulation for the rapid and gradual tasks (Audio * Task interaction, p<.001). While the psychophysical results are consistent with auditory stimulation increasing the attentional strobe rate, the MEG analysis demonstrated such changes may be driven through shifts in theta power.

The ability to imagine is arguably one of our most important human cognitive faculties, with far reaching importance for perception, sensory simulation (future & past) and maybe even creating new inventions. However, imagery strength displays inexplicable large idiosyncrasies across the population. To investigate a possible cause of these idiosyncrasies we studied the role of cortical excitability in two areas of the brain often implicated in mental imagery: visual and pre-frontal cortex. To measure imagery strength we used the binocular rivalry technique. In this task participants were required to imagine one of two patterns followed by a brief binocular rivalry display, with imagery strength measured by percent of trials primed. This method has been shown to give a reliable and indirect measure of the sensory strength of visual imagery. We found a significant negative correlation between resting state visual cortex excitability (measured by TMS and fMRI) and imagery strength. Additionally, using tDCS we show a causal effect of neural excitability on imagery strength, decreasing visual cortex excitability leads to stronger imagery, whereas decreasing excitability resulted in a slight decrease in imagery strength. Conversely we found the opposite pattern for the pre-frontal cortex, cortical excitability (measured by fMRI) positively correlated with imagery strength and increasing pre-frontal cortex excitability boosted imagery strength. We propose that the strength of visual imagery is largely dictated by the ratio of top-down signal strength to visual cortex neural noise, with optimal image generation occurring with strong top-down signals and low-levels of noise in the sensory cortices.

Vision can be considered as a process of probabilistic inference. In a Bayesian framework, perceptual estimates from sensory information are combined with prior knowledge, with a stronger influence of the prior when the sensory evidence is less certain. Here, we explored the behavioural and neural correlates of such a prior in the context of orientation processing. First, we asked participants (all n≥5) to judge whether a stimulus was oriented closer to vertical or oblique for two stimulus types (spatially-filtered noise textures, sinusoidal gratings) and three manipulations of certainty (orientation bandwidth, contrast, duration). We found participants consistently had a bias towards reporting oblique orientations during conditions of high certainty but that this bias was reduced when sensory evidence was less certain. Second, we measured (n=8) event-related fMRI BOLD responses in human primary visual cortex (V1) and manipulated certainty via stimulus contrast (100% vs. 3%). We observed a higher response to oblique orientations for high contrast stimuli that reversed to reveal a vertical bias at low contrast. We then trained a multivariate classifier on the pattern of responses in V1 to cardinal and oblique orientations. We found that the classifier categorized a wider range of orientations as cardinal for low contrast stimuli. These findings are consistent with the operation of a prior in primary visual cortex during processing of stimulus orientation, biasing perception towards those orientations that are more prevalent in the natural environment.

Adaptive behaviour often requires us to divide attention across information sources. Surprisingly, recent data (Yamani et al., 2015, ASP) have suggested that the ability to process multiple visual signals simultaneously improves under high perceptual demand; parallel processing efficiency, as measured by the workload capacity coefficient C(t) (Townsend & Eidels, 2011, PBR), was higher in cluttered than in uncluttered displays. The current work asked whether the effects of clutter on C(t) were driven by bottom-up or top-down attentional processes.

Participants performed a speeded letter identification task in a redundant-targets paradigm. In the uncluttered condition, red or green target(s) appeared alone in the display. In the pop-out condition, the target(s) appeared amongst distractors of the alternative colour. In the uniform condition, the target(s) appeared amongst distractors of the same colour. The uncluttered and pop-out conditions thus allowed efficient bottom-up target selection, while the uniform condition required top-down selection. C(t) was calculated from RT distributions for single- and redundant-target trials. Higher values indicate more efficient processing on redundant-target trials.

RTs were shorter for redundant-target than for single-target trials, confirming that redundant targets were processed in parallel. And as expected, RTs were shorter in the uncluttered and pop-out conditions than in the uniform condition, confirming that bottom-up processes allowed faster target selection. C(t), however, was greatest in the uniform condition, where top-down control was necessary to select target(s) from distractors. Results suggest that bottom-up processes reduce the efficiency of parallel target processing, perhaps by producing competition for selection between targets.