Time to Decide. (360G-Wellcome-106988_Z_15_Z)

The objective of this proposal is to investigate the temporal dynamics of simple perceptual decisions in Drosophila, with a view toward uncovering general mechanisms of neural information processing at timescales from hundreds of milliseconds to several seconds. In his classical essay The Problem of Serial Order in Behavior, the psychologist Karl Lashley emphasized the ubiquity of brain processes that unfold over time: Temporal integration is not found exclusively in language; the coordination of leg movements in insects, the song of birds, the control of trotting and pacing in a gaited horse, the rat running the maze, the architect designing a house, and the carpenter sawing a board present a problem of sequences of action which cannot be explained in terms of successions of external stimuli. In spite of the ubiquity of the problem, there have been almost no attempts to develop physiological theories to meet it. Although Lashley wrote this passage more than 60 years ago, the fundame ntal problem of how activity sequences are generated remains largely unsolved. Temporal processing is integral also to decision-making because the information necessary to commit to a choice is rarely available all at once but must be gathered over time. A large literature, whose beginnings stretch back to the 19th century, documents systematic variations in the speed of perceptual judgments with stimulus strength: easy decisions, based on strong, unambiguous sensory data, tend to be fast; di fficult decisions, based on weak or conflicting data, tend to be slow. This difficulty-dependent cost of decision time is thought to reflect an underlying need to construct time-averaged sensory representations. Just like engineers average signals over time to reduce the effects of contaminating noise, the brain appears to improve its signal-to-noise ratio by integrating information from sequential samples. The duration of the integration period depends on the quality of the sensory data and the desired response accuracy or confidence. Beautiful as these ideas are, the neural mechanisms that allow neurons to accumulate information, compare the accumulated signal to a response criterion, and discharge behaviour when the criterion is met remain elusive. The complexity of vertebrate brains, the difficulty of molecular interventions, and the effort required to generate and analyse genetic variants have presented serious barriers to progress. These barriers have begun to ero de with our recent discovery that fruit flies, like mammals, take longer to commit to difficult perceptual choices than to easy ones, and that quantitative relationships link speed, accuracy, and task difficulty. The exact mathematical form of these relationships is predicted by integrator models that were originally formulated to describe human behaviour. Drosophila thus appears to accumulate sensory data in the lead-up to a choice, making this type of temporal processing amenable to genetic di ssection. A small screen of candidate genes uncovered an unexpected role for FoxP in decision-making. FoxP mutants are slower to commit than wild-type flies, and, despite taking longer, are also more error-prone - precisely the constellation of symptoms one might expect in an a