78

No-Report Paradigm in Binocular Rivalry The stronger frontal correlations with transitions in binocular rivalry may have been due to the fact that the cognitive processes involved in deciding what to report involve frontal activations. That ideamotivatedtheno-reportparadigm.Ifoneeyeisshownagratingmovingtotheleftandtheother eye is shown a grating moving to the right, the subject is aware of leftward motion (usually in the wholevisual field), then rightward motion, then leftward motion, etc. Using a method first discovered in [15], a recent study used a characteristic eye movement called optokinetic nystagmus (OKN) that correlates with the perceived direction of motion as indexed by the subjects’ reports [16]. Smooth pursuit to the left plus sharp jerky motions to the right correlate with reports of conscious perception of the grating as moving leftward and there is a corresponding association for rightward motion. In binocular rivalry, there are always brief intermediate states that involve patches of percepts of the two stimuli and aspects of the nystagmus correlated with that too. Once the researchers had verified the accuracy of nystagmus using self-reports, they put subjects in the scanner. However, once in the scanner, subjects were not given any task and they were not asked for any reports. The researchers then looked at differences in brain activations when nystagmus indicatedaperceptualtransition.Ofcourse,subjectscouldtestifyaftertheexperimentwasoverthattheir percepts were alternating as usual. The transitions mainly reflected differences in perceptual areas in the back and middle of the head. The article summarizes: ‘Importantly, when observers passively experienced rivalry without reporting perceptual alternations, a different picture [i.e., different from what happens with report] emerged’. That different picture is that differential neural activity 1006 Trends in Cognitive Sciences, December 2019, Vol. 23, No. 12Trends in Cognitive Sciences in prefrontal areas was minorcomparedwithactivityintemporalandparietalregions.They‘conclude that prefrontal areas are associated with active report and introspection’ ([16], see p. 1738). The article’s title reflects this emphasis: ‘Binocular Rivalry: Frontal Activity Relatesto Introspection andAction but Not to Perception’. Thisresultledtoaflurryofcontroversy[17,18]inwhichdifferenttypesofexperimentsseemedtodiffer in whether they showedprefrontal differences in perceptual transitions in a paradigm closely related to binocular rivalry. These results presented serious challenges to the conclusion just described. In particular, prefrontal reflections of perceptual contents were decoded in a no-report experiment with monkeys using electrophysiological methods (electrodes inserted in cortical regions) that are known to be more sensitive to neural activations than the fMRI used in the result mentioned in the last paragraph [17]. Impressively, these resultsusedmonkeysthathadnotbeentrainedonadiscrimination task, ruling out covert decision-making that would have been expected to make a prefrontal difference [T. Panagiotaropoulos (corresponding author of [17]), personal communication]. Later work in the same laboratory has been devoted to recording from grids of microelectrodes (‘Utah’ arrays) placed in the prefrontal cortex of monkeys inabinocular rivalry setup withgratings in different directions, suggesting decoding of direction from the prefrontal cortex [19,20]. Inaddition,another group was able to decode perceptual contents from prefrontal areas in binocular rivalry [18]. However, this result did not use a no-report paradigm so the prefrontal representation could have been linked to the cognitive processes underlying reporting. Acaution about decoding from prefrontal cortex: because of linkages across the whole brain, it may bepossibleto‘decode’anythingfromanywhereinthebrainifonehassufficientlysensitivedetection. The real issue is which regions involve optimal decoding over a wide range of circumstances. In sum, work using the no-report paradigm has shown that even without reports, both contents and transitions can be decoded from the frontal cortex during binocular rivalry. However, the no-report paradigm is not dead. It can be modified to avoid the problems just described, as I argue in the next section. The Need for a No-Post-perceptual Cognition Paradigm Suppose it is confirmed using microelectrode arrays that conscious perceptions (not just transitions) can be decoded from the prefrontal cortex in binocular rivalry in a no-report paradigm. Would that result showthattheprefrontal cortex is part ofthe neural basisofconsciousperception?No,because of the ‘bored monkey’ problem. The monkeys in these experiments spend hours looking at gratings going up and down without any task other than fixating. If you were in this perceptual situation, you mighthavesomecognitivestates–thinking,wondering,questioning,musing,andthelike–concerningthegratingmovingupwhenitismovingupandthegratingmovingdownwhenitismovingdown. Asaconsequence,themicroelectrodearrayscouldbetappingpost-perceptualcognitiveprocessing concerning which way the gratings are moving rather than the perceptions of the gratings themselves. Note that I am not talking about daydreaming or mind-wandering. Those states cannot be controlled and, in any case, would not engender systematic error. The problem is that the monkeys may be undergoing cognitive processes that are systematically aligned with one or another of the rivalrous percepts. This problem reveals a flaw in the reasoning behind the no-report paradigm. Eliminating report is successful in isolating the neural basis of conscious perception only if it eliminates post-perceptual cognitive processing such as thought and judgment about the reportable properties that is systematically correlated with one of the perceptual representations. Thesameprobleminfects recently rediscovered [21] older versions of the no-report paradigm involving humans [22] since nothing was done to prevent subjects thinking about their alternating percepts. So, what we really need to do is to replace the no-report paradigm with a ‘no-cognition’ paradigmor better, a no-differential (between rivalrous percepts)-post-perceptual cognition paradigm (to Trends in Cognitive Sciences, December 2019, Vol. 23, No. 12 1007Trends in Cognitive Sciences Le eye Right eye Different colors Figure 2. Inconspicuous Binocular Rivalry. Quasi-random motion stimuli used in binocular rivalry experiments by [23]. The alternation between theblueandreddotsatthetopisverynoticeable but the alternation of the same-color dots is not because the dot pattern is everchanging. Thanks to Jan Brascamp for this figure. avoid begging the question against views that take conscious perception to be cognitive). However, this may seem manifestly impossible. When subjects see things, they are free to make perceptual judgments and think about what they see. You cannot stop subjects– including monkeys– thinking. Weseemto be at an impasse.