During perceptual decisions, subjects often rely more strongly on early, rather than late, sensory evidence, even in tasks when both are equally informative about the correct decision. This early psychophysical weighting has been explained by an integration-to-bound decision process, in which the stimulus is ignored after the accumulated evidence reaches a certain bound, or confidence level. Here, we derive predictions about how the average temporal weighting of the evidence depends on a subject's decision confidence in this model. To test these predictions empirically, we devised a method to infer decision confidence from pupil size in 2 male monkeys performing a disparity discrimination task. Our animals' data confirmed the integration-to-bound predictions, with different internal decision bounds and different levels of correlation between pupil size and decision confidence accounting for differences between animals. However, the data were less compatible with two alternative accounts for early psychophysical weighting: attractor dynamics either within the decision area or due to feedback to sensory areas, or a feedforward account due to neuronal response adaptation. This approach also opens the door to using confidence more broadly when studying the neural basis of decision making. An animal's ability to adjust decisions based on its level of confidence, sometimes referred to as "metacognition," has generated substantial interest in neuroscience. Here, we show how measurements of pupil diameter in macaques can be used to infer their confidence. This technique opens the door to more neurophysiological studies of confidence because it eliminates the need for training on behavioral paradigms to evaluate confidence. We then use this technique to test predictions from competing explanations of why subjects in perceptual decision making often rely more strongly on early evidence: the way in which the strength of this effect should depend on a subject's decision confidence. We find that a bounded decision formation process best explains our empirical data.

Fine judgments of stereoscopic depth rely mainly on relative judgments of depth (relative binocular disparity) between objects, rather than judgments of the distance to where the eyes are fixating (absolute disparity). In macaques, visual area V2 is the earliest site in the visual processing hierarchy for which neurons selective for relative disparity have been observed (Thomas et al., 2002). Here, we found that, in macaques trained to perform a fine disparity discrimination task, disparity-selective neurons in V2 were highly selective for the task, and their activity correlated with the animals' perceptual decisions (unexplained by the stimulus). This may partially explain similar correlations reported in downstream areas. Although compatible with a perceptual role of these neurons for the task, the interpretation of such decision-related activity is complicated by the effects of interneuronal "noise" correlations between sensory neurons. Recent work has developed simple predictions to differentiate decoding schemes (Pitkow et al., 2015) without needing measures of noise correlations, and found that data from early sensory areas were compatible with optimal linear readout of populations with informationlimiting correlations. In contrast, our data here deviated significantly from these predictions. We additionally tested this prediction for previously reported results of decision-related activity in V2 for a related task, coarse disparity discrimination (Nienborg and Cumming, 2006), thought to rely on absolute disparity. Although these data followed the predicted pattern, they violated the prediction quantitatively. This suggests that optimal linear decoding of sensory signals is not generally a good predictor of behavior in simple perceptual tasks.

Although luminance and color are thought to be processed independently at early stages of visual processing, there is evidence that they interact at later stages. For example, chromatic information has been shown to enhance or suppress depth from luminance depending on whether chromatic edges are aligned or orthogonal with luminance edges. Here we explored more generally how chromatic information interacts with luminance information that specifies shape from shading. Using a depth-matching task, we measured perceived depth in sinusoidal and square-wave gratings (specifying close-to sinusoidal and triangle-wave depth profiles, respectively) in three conditions. In the first, as we varied luminance contrast in the presence of an orthogonal chromatic grating, perceived depth increased (consistent with classical shape from shading). When we held the luminance at a fixed contrast and varied the chromatic grating in the other two conditions (orthogonal or aligned), we found large and inconsistent individual differences. Some participants exhibited the expected pattern of enhancement and suppression, but most did not, either for the sinusoidal or square-wave stimuli. Our results cast doubt on the idea that the interaction demonstrates a single high-level heuristic linked to depth perception. Instead, we speculate that interactions are more likely due to early cross-channel masking.

During perceptual decisions subjects often rely more strongly on early rather than late 21 sensory evidence even in tasks when both are equally informative about the correct 22 decision. This early psychophysical weighting has been explained by an integration-to-23 bound decision process, in which the stimulus is ignored after the accumulated evidence 24 reaches a certain bound, or confidence level. Here, we derive predictions about how the 25 average temporal weighting of the evidence depends on a subject's decision-confidence 26 in this model. To test these predictions empirically, we devised a method to infer 27 decision-confidence from pupil size in monkeys performing a disparity discrimination 28 task. Our animals' data confirmed the integration-to-bound predictions, with different 29 internal decision-bounds accounting for differences between animals. However, the data 30 could not be explained by two alternative accounts for early psychophysical weighting: 31 attractor dynamics either within the decision area or due to feedback to sensory areas, or 32 a feedforward account due to neuronal response adaptation. This approach also opens 33 the door to using confidence more broadly when studying the neural basis of decision-34 making. 35 36 37 measuring the animal's subjective decision confidence. 54 55Measuring decision confidence psychophysically is relatively difficult, particularly in animals, and 56 increases the complexity of a task, as e.g. for post-decision wagering 12,13 , hence requiring 57 additional training. To avoid these difficulties we devised a metric based on the monkeys' pupil 58 size. Combining this metric for decision confidence with psychophysical reverse correlation 3,14,15 59 allowed us to quantify the animals' psychophysical weighting strategy for different levels of 60 inferred decision-confidence, and test our model predictions. The animals showed clear early 61 psychophysical weighting on average. But separating this analysis by inferred decision 62 confidence revealed that early psychophysical weighting was largely restricted to high 63 confidence trials. In fact, on low inferred confidence trials the animals weighted the stimulus 64 relatively uniformly or even slightly more towards the end of the trial. Such behavior matched 65 the predictions of the integration-to-bound model. Furthermore, the differences between both 66 animals could be accounted for by the model by differences in the only free parameter -their 67 internal decision-bound. 68 69In contrast, the animals' behavior could not be fully explained by two alternative accounts of 70 early psychophysical weighting. The first alternative account are models in which the decision-71 3 stage provides self-reinforcing feedback to the sensory neurons 16 , as suggested, e.g. for 72 probabilistic inference 17 , or by attractor dynamics within the decision-making area 28 . The 73 second, recent alternative proposal is that the early weighting simply reflects the feed-forward 74 effect of the dynamics (gain control or adaptation) o...

Highlights 18• We developed an approach to train macaque monkeys head-free on visuomotor tasks 19 requiring measurements of eye position 20• The setup is inexpensive, easy to build, and readily adjusted to the animal without the need 21 for sedation 22 Abstract 31We describe a modified system for training macaque monkeys without invasive head 32 immobilization on visuomotor tasks requiring the control of eye-movements. The system 33 combines a conventional primate chair, a chair-mounted infrared camera for measuring 34 eye-movements and a custom-made concave reward-delivery spout firmly attached to 35 the chair. The animal was seated head-free inside the chair but the concavity of the spout 36 stabilized its head during task performance. Training on visual fixation and 37 discrimination tasks was successfully performed with this system. Eye-measurements, 38 such as fixation-precision, pupil size as well as micro-saccades were comparable to 39 those obtained using conventional invasive head-fixation methods. The system is 40 inexpensive (~$40 USD material cost), easy to fabricate in a workshop (technical 41 drawings are included), and readily adjustable between animals without the need to 42 immobilize or sedate them for these adjustments. 43 44 45