Most models assume that top-down attention enhances the gain of sensory neurons tuned to behaviorally-relevant stimuli (on-target gain). However, theoretical work suggests that when targets and distracters are highly similar, attention should enhance the gain of neurons that are tuned away from the target, because these neurons better discriminate neighboring features (off-target gain). While it is established that off-target neurons support difficult fine discriminations, it is unclear if top-down attentional gain can be optimally applied to informative off-target sensory neurons or if gain is always applied to on-target neurons, irrespective of task demands. To test the optimality of attentional gain in human visual cortex, we used fMRI and an encoding model to estimate the response profile across a set of hypothetical orientation-selective channels during a difficult discrimination task. The results suggest that top-down attention can adaptively modulate off-target neural populations, but only when the discriminanda are precisely specified in advance. Furthermore, logistic regression revealed that activation levels in off-target orientation channels predicted behavioral accuracy on a trial-by-trial basis. Overall, these data suggest that attention does not always increase the gain of sensory-evoked responses, but instead may bias population response profiles in an optimal manner that respects both the tuning properties of sensory neurons and the physical characteristics of the stimulus array.

Selective attention supports the prioritized processing of relevant sensory information to facilitate goal-directed behavior. Studies in human subjects demonstrate that attentional gain of cortical responses can sufficiently account for attention-related improvements in behavior. On the other hand, studies using highly trained nonhuman primates suggest that reductions in neural noise can better explain attentional facilitation of behavior. Given the importance of selective information processing in nearly all domains of cognition, we sought to reconcile these competing accounts by testing the hypothesis that extensive behavioral training alters the neural mechanisms that support selective attention. We tested this hypothesis using electroencephalography (EEG) to measure stimulus-evoked visual responses from human subjects while they performed a selective spatial attention task over the course of~1 month. Early in training, spatial attention led to an increase in the gain of stimulusevoked visual responses. Gain was apparent within~100 ms of stimulus onset, and a quantitative model based on signal detection theory (SDT) successfully linked the magnitude of this gain modulation to attention-related improvements in behavior. However, after extensive training, this early attentional gain was eliminated even though there were still substantial attention-related improvements in behavior. Accordingly, the SDT-based model required noise reduction to account for the link between the stimulus-evoked visual responses and attentional modulations of behavior. These findings suggest that training can lead to fundamental changes in the way attention alters the early cortical responses that support selective information processing. Moreover, these data facilitate the translation of results across different species and across experimental procedures that employ different behavioral training regimes. Author summarySelective attention can enhance processing of sensory information via sensory gain and neural noise reduction. However, the extent to which these 2 mechanisms contribute to improvement in perceptual performance during attention is still debated. We hypothesized PLOS Biology | https://doi.org/10.1371/journal.pbio

Here, we review the role of top-down attention in both the acquisition and the expression of perceptual learning, as well as the role of learning in more efficiently guiding attentional modulations. Although attention often mediates learning at the outset of training, many of the characteristic behavioral and neural changes associated with learning can be observed even when stimuli are task irrelevant and ignored. However, depending on task demands, attention can override the effects of perceptual learning, suggesting that even if top-down factors are not strictly necessary to observe learning, they play a critical role in determining how learning-related changes in behavior and neural activity are ultimately expressed. In turn, training may also act to optimize the effectiveness of top-down attentional control by improving the efficiency of sensory gain modulations, regulating intrinsic noise, and altering the read-out of sensory information.

Learning to better discriminate a specific visual feature (i.e., a specific orientation in a specific region of space) has been associated with plasticity in early visual areas (sensory modulation) and with improvements in the transmission of sensory information from early visual areas to downstream sensorimotor and decision regions (enhanced readout). However, in many real-world scenarios that require perceptual expertise, observers need to efficiently process numerous exemplars from a broad stimulus class as opposed to just a single stimulus feature. Some previous data suggest that perceptual learning leads to highly specific neural modulations that support the discrimination of specific trained features. However, the extent to which perceptual learning acts to improve the discriminability of a broad class of stimuli via the modulation of sensory responses in human visual cortex remains largely unknown. Here, we used functional MRI and a multivariate analysis method to reconstruct orientation-selective response profiles based on activation patterns in the early visual cortex before and after subjects learned to discriminate small offsets in a set of grating stimuli that were rendered in one of nine possible orientations. Behavioral performance improved across 10 training sessions, and there was a training-related increase in the amplitude of orientation-selective response profiles in V1, V2, and V3 when orientation was task relevant compared with when it was task irrelevant. These results suggest that generalized perceptual learning can lead to modified responses in the early visual cortex in a manner that is suitable for supporting improved discriminability of stimuli drawn from a large set of exemplars.

words)Attention supports the selection of relevant sensory information from competing irrelevant sensory information. This selective processing is thought to be supported via the attentional gain amplification of sensory responses evoked by attended compared to unattended stimuli. However, recent studies in highly trained subjects suggest that attentional gain plays a relatively modest role and that other types of neural modulations -such as a reduction in neural noise -better explain attention-related changes in behavior. We hypothesized that the amount of training may alter neural mechanisms that support attentional selection in visual cortex. To test this hypothesis, we investigated the influence of training on attentional modulations of stimulus-evoked visual responses by recording electroencephalography (EEG) from humans performing a selective visuospatial attention task over the course of one month. Early in training, visuospatial attention induced a robust attentional gain amplification of sensory-evoked responses in contralateral visual cortex that emerged within ~100ms after stimulus onset, and a quantitative model based on signal detection theory (SDT) successfully linked this attentional gain amplification to attention-related improvements in behavior. However, after training, this attentional gain amplification of visual responses was almost completely eliminated and modeling suggested that noise reduction was required to link the amplitude of visual responses with attentional modulations of behavior. These findings suggest that the neural mechanisms supporting selective attention can change as a function of training and expertise, and help to bridge different results from studies carried out in different model systems that require substantially different amount of training.