Behavioral and Brain Measures of Phasic Alerting Effects on Visual Attention

In the present study, we investigated effects of phasic alerting on visual attention in a partial report task, in which half of the displays were preceded by an auditory warning cue. Based on the computational Theory of Visual Attention (TVA), we estimated parameters of spatial and non-spatial aspects of visual attention and measured event-related lateralizations (ERLs) over visual processing areas. We found that the TVA parameter sensory effectiveness a, which is thought to reflect visual processing capacity, significantly increased with phasic alerting. By contrast, the distribution of visual processing resources according to task relevance and spatial position, as quantified in parameters top-down control α and spatial bias windex, was not modulated by phasic alerting. On the electrophysiological level, the latencies of ERLs in response to the task displays were reduced following the warning cue. These results suggest that phasic alerting facilitates visual processing in a general, unselective manner and that this effect originates in early stages of visual information processing.


ERL analyses following the jackknifing procedure
We repeated analyses on ERL latency measures following the jackknifing procedure introduced by Miller et al., 1998 (see also Kiesel et al., 2008;Ulrich & Miller 2001), an alternative measure of the timing of ERPs, assumed to be less prone to noise-related distortions than single-participant measurements. With the jackknife approach, latencies are measured from n grand averages, computed from subsamples of n-1, with each participant being omitted from one of the subsample grand averages. We analyzed i) peak latencies and ii) fractional peak latencies (i.e., the time point where the voltage reached 80% of the maximum peak) in the time windows 120-210 for unilateral displays and 180-290 for bilateral displays.
In the analyses on peak latencies after jackknifing, the same descriptive pattern of results was found as reported in the main article (see Table below The additional analyses support the critical finding of a cue-related reduction in ERL latencies which we found in the latency measures reported in the main article. They do not confirm the interaction between cue effects and conditions, suggesting that noise contributing to the variance in ERL latency measures affected the cue-related effect in the varying display conditions differently.

Table
Mean and standard errors of the mean of the ERL peak and onset latencies measured in four display conditions of the partial report task (1T: single target letter, 2T ipsi: target plus second target in the ipsilateral hemifield, TD ipsi: target plus distractor in ipsilateral, TD contra: target plus distracter in contralateral hemifield), separately for trials with (cue) and without an alerting tone (no cue) following the jackknife-procedure (Miller et al., 1998

Correlation analyses
We found a significant correlation (Pearson) between the cue-related increase in parameter sensory effectiveness a and the cue-related ERL latency reduction in the single target condition, but not with ERL latency reductions in the other conditions (see Figure 1). We further explored whether cueeffects on the other parameters top-down control α and spatial bias w index may have masked the correlations between cue effects on parameter a and ERL latencies in the target-distracter and dual target conditions. This does not seem to be the case: Partial correlations including cue-effects on α and w index as covariates did not change the pattern of results. The partial correlations between the effects of cue on sensory effectiveness a and ERL latency reduction in the single target condition One participant had overall larger a-values and also showed a stronger cue-effect on a (Figure 1).

Non-parametric (Spearman Rho) correlation analyses between the cue-related increase in parameter
sensory effectiveness a and the cue-related ERL latency did not approach significance in the single Finally, we explored the relationship between baseline performance (report accuracy in the single target condition) and the cue effect on parameter sensory effectiveness a. We found a marginal significant positive correlation between the individual alerting effect on a and baseline accuracy [parametric (Pearson): r=.43,p=.08;p=.06], reflecting that participants with lower baseline performance tended to benefit less from the cue. It would be interesting to further explore this relationship in future studies designed to systematically test the influence of baseline performance on phasic alerting. Notably, in this study individual differences in baseline performance were partly cancelled out by individually adjusting the exposure durations. Figure 1: Correlations between the cuerelated increase in sensory effectiveness (a Cuea NoCue ) and the relative reduction of ERL latencies by phasic alerting (ERL NoCue -ERL Cue ) in the four display conditions (1T: single-target letter alone, 2T ipsi: target plus second target in ipsilateral hemifield, TD ipsi: target plus distractor in ipsilateral hemifield, and TD contra: target plus distractor in the contralateral hemifield.

Figure 2:
Correlations between the cuerelated increase in sensory effectiveness (a Cuea NoCue ) and baseline performance in the single target condition.