Distributed range adaptation in human parietal encoding of numbers
The brain represents magnitudes through the collective activity of neural populations, whose non-monotonic tuning properties determine the nature and precision of the population neural code. Whether and how this code adapts to changes in the statistics of the encoded magnitudes remains unknown. Here we probe the adaptation of the encoding of numbers in human parietal cortex, using functional MRI during a numerosity-estimation task in which we vary the range of possible numbers. Tracking the tuning properties of number-sensitive populations as the range changes, we show that their receptive fields shift and scale in adaptation to the range, following a structured and predictable pattern. This distributed range adaptation implements efficient coding dynamically: the resulting precision of the neural code varies with the range and is accompanied by corresponding changes in behavioral precision. At the participant level, the degree of neural tuning adaptation significantly correlates with the change in behavioral variability. Our results extend static sensory efficient coding to the adaptive representation of abstract magnitudes, via a neural mechanism of distributed range adaptation that may be a canonical property of neural encoding circuits.