Attention III

Measuring the frame rate of visual perception - the role of brain oscillations in the temporal segregation of visual events

Andreas Wutz, Nathan Weisz, Christoph Braun and David Melcher

Despite phenomenologically continuous, increasing evidence suggests visual perception to be discrete and points to the outstanding role of ongoing brain oscillations. We addressed this empirical question using a forward masking procedure with variable stimulus onset asynchrony (SOA), while the observers were engaged in an enumeration task. Concurrently we recorded electrophysiological brain activity using MEG. Enumeration is a particularly interesting task in this context, as a small number of items (up to four) is supposed to be simultaneously apprehended in one instant – a phenomenon called subitizing. The to be enumerated items were superposed upon the masking pattern, so that the performance melioration with increasing mask-target SOA in this task supposedly depended upon the observer’s accumulative ability to segregate the temporal onset of mask and target displays. We hypothesize that correct trials indicate temporal segregation, while incorrect trials are associated with temporal integration of mask and target information. We searched for power and phase differences between these proportions of trials at pre- and peri-mask intervals and found evidence supporting the idea that oscillatory brain activity could mediate between temporal integration and segregation of visual events and therefore serve as a frame rate of visual perception.

A computational model of microsaccadic responses to shifts of covert attention

Ralf Engbert

Even when we fixate an object of interest, our eyes perform miniature (or fixational) eye movements, which are classified as slow movements (physiological drift) and microsaccades. Suddenly upcoming stimuli in the periphery induce covert attentional shifts which are followed by a saccadic eye movements later on. Before the saccadic response, however, microsaccades are modulated in terms of their rate-of-occurrence as well as their spatial orientations. I present a computational model that integrates lower-level perceptual responses and higher-level attentional signal to reproduce behavioral findings on fixational eye movements. Microsaccadic rate modulations as well as the temporal variation of the orientation bias of microsaccades can be predicted from the model. Using extensive numerical simulations, it turned out that the type of the attentional cue (endogeneous or exogeneous) and the temporal delay between perceptual onset and the resulting attentional shift are critical for the pattern of results observed in experiments. Results are relevant to the more the general problem of integration of low-level perception and top-down cognitive processes in sensorimotor control.

Perceptual learning at the location of predictive remapping

Martin Rolfs, Nick Murray-Smith and Marisa Carrasco

Just before the eyes move, predictive remapping of visual attention facilitates perception at future relevant locations [Rolfs et al., 2011, Nature Neuroscience, 14, 252–256], providing a functional account of anticipatory responses in many retinotopic brain areas when an imminent saccade will bring a stimulus into a neuron's receptive field [Duhamel et al., 1992, Science, 255, 90–92]. Here, we examine the impact of predictive remapping on visual processing using a perceptual learning protocol. Observers performed a fine orientation discrimination task (±5º off a predefined reference), judging a peripheral Gabor embedded in noise. Over five days of training, observers made a saccade following a stereotypic step of the fixation spot while the Gabor always appeared just before the eye movement. Contrast thresholds decreased continuously—the signature of learning. During the final three days, observers maintained fixation and we tested whether learning transferred to (1) the remapped location of the training stimulus (offset from training location by the inverted saccade vector), or (2) a spatially matched control location (same distance as remapped location from training location). Observers’ thresholds at the remapped location were significantly lower than at the control location. This enhanced visual sensitivity indicates that learning transferred under predictive remapping.

Trans-saccadic priming of location and orientation

Kay Ritchie, Amelia Hunt and Arash Sahraie

We experience visual stability despite shifts in the retinal array produced by saccades, but the mechanisms supporting stability are the subject of some debate. We have previously shown that a pre-saccadic prime in the blind visual field of hemianopic participants leads to increased detection and discrimination in that location when it is brought into the sighted field by an eye movement. Here we replicated this finding in neurologically intact controls using briefly-presented pre-saccadic primes. Discrimination of post-saccadic targets was more accurate when they were presented in the same location as the prime than when there was no prime or the prime was in a different location. In our second experiment we examined the effect of the orientation of the pre-saccadic prime on discrimination of the orientation of the post-saccadic target. Performance was significantly improved when the prime was presented in the same orientation as the target. This orientation-based facilitation was limited to targets in the same retinotopic location as the target, and to a lesser extent, the same spatiotopic location as the target. The results suggest that location and, to a lesser extent, orientation information are maintained in spatiotopic coordinates when saccades shift visual input across the retina.

How does memory search interact with the attentional blink?

Trafton Drew, Ashley M. Sherman and Jeremy M. Wolfe

Humans are remarkably good at identifying target objects in Rapid Serial Visual Presentation (RSVP) sequences of objects. However, suppose observers must hold several possible targets in memory (Monitor the stream for cow, spoon, pen, or key). This necessitates hybrid search of memory and of the RSVP stream. In visual search tasks, the amount time needed to evaluate each item increases linearly with the log of the number of possible targets (Wolfe, in press, Psychological Science). In RSVP, we have shown that the presentation rate required to reach a criterion target detection rate increases as a function of the log of the number of possible targets (Drew & Wolfe, 2012, VSS). Does memory search for a match to the current item in an RSVP stream influence processing of subsequent items in the stream? To answer this, we used a modified attentional blink paradigm. T1 was an object that needed to be matched to the memory set. If T2 was an object, the magnitude of the ‘blink’ produced by T1 was strongly dependent on memory set size. However, if T1 was an object and T2 a number, T1 produced the same magnitude blink regardless of memory set size.

Differential involvement of dorsolateral prefrontal cortex in difficult visual search: a TMS study

Gorana Pobric and Johan Hulleman

Recently, Young and Hulleman (JEP:HPP, in press) proposed a new theoretical framework for visual search. The framework models visual search as a sequence of fixations where items are processed in parallel within a fixation. Crucially, the difficulty of the search task determines the number of items processed per fixation. If the search task is easy, many items will be processed, but when the search task is very difficult only a single item will be processed. We tested this assumption with transcranial magnetic stimulation (TMS). Participants searched for T amongst L (medium search) or for a square with a smaller square in the left top corner amongst squares with a smaller square in one of the other corners (difficult search). Display size was manipulated to equate tasks for search times. Low frequency repetitive TMS (1Hz for 10 min at 64% of machine output) was applied offline over the right frontoparietal network. Performance was selectively impaired for difficult search after dorsolateral prefrontal cortex (DLPFC) stimulation, suggesting a crucial involvement in the precise saccadic targeting and increased attentional control demanded by difficult search. This finding supports fundamental distinction between very difficult search and easier search made by the framework.

Confusion: some strategies to prevent successful object tracking

Nick Scott-Samuel, Gavin Holmes, Roland Baddeley and Innes Cuthill

What factors influence one’s ability to track a target object moving amongst dynamic distractor items? Biologists hypothesise a “confusion effect”: the idea that predation success decreases as prey group size increases. This concept resonates with the human visual search literature, and has implications for the general question of how to hide in a crowd. We used a novel version of a standard multiple object tracking display. Subjects tracked one object amongst several, and errors in locating the final position of that object were recorded. The predictability of all objects’ motion paths was systematically varied, as was their number. In Experiment 1, we compared the effect of using background-matching and two-dimensional “dazzle” (high contrast binary) textures on plain or textured backgrounds. In Experiment 2, we manipulated number and area. Subjects were at ceiling for stimuli with relatively predictable paths. At higher levels of unpredictability a confusion effect appeared, and then increased with decreasing path predictability. This increase was greater for larger numbers of objects. Surprisingly, varying the textures of objects and background had no effect. Finally, density, not number, influenced error rates. So for optimal confusion, don’t bother with colouration; the critical factors are the unpredictability and density of your group.

Stimulation of the left parietal lobe improves sustained attention in right parietal lobe patients: tipping the inter-hemispheric balance with TMS

Sara Agosta, Florian Herpich and Lorella Battelli

Right parietal patients are impaired at performing visual tracking of multiple objects. In particular they have difficulties in tracking moving stimuli in the left visual field during simultaneous presentation of another similar target in the opposite hemifield, a deficit likely caused by extinction (Battelli et al., 2001). One of the hypothesis is (Kinsbourne, 1977) that extinction is a consequence of an increased inhibition exerted on the damaged hemisphere by the hyperactive unaffected hemisphere. Here we used transcranial magnetic stimulation (TMS) on the healthy hemisphere to tip the balance between hemispheres and relieve extinction symptoms. Methods: Patients underwent two counterbalanced sessions: low frequency TMS over the left (healthy) parietal lobe and sham control stimulation. We compared their performance before and after TMS. Subjects were asked to track 2 (unilateral condition) or 4 (bilateral condition) moving discs amidst moving distracters (8 total discs, four in each hemifield) in the left and/or right hemifield. Results: Patients’ performance improved immediately after TMS and continued to improve in the next 30 minutes for both unilateral and bilateral conditions. Conclusion: TMS might have beneficial effects on rehabilitation of extinction re-balancing the activity of the two parietal lobes by suppressing the inhibiting activity of the unaffected hemisphere.