Whether the human brain is equipped with a special neural substrate for numbers, or rather with a common neural substrate for processing of several types of magnitudes, has been the topic of a long-standing debate. The present study addressed this question by using functional magnetic resonance imaging (fMRI) and event-related potentials (ERPs) together with the size-congruity paradigm, a Stroop-like task in which numerical values and physical sizes were varied independently. In the fMRI experiment, a region-of-interest analysis of the primary motor cortex revealed interference effects in the hemisphere ipsilateral to the response hand, indicating that the stimulus-stimulus conflict between numerical and physical magnitude is not completely resolved until response initiation. This result supports the assumption of distinct comparison mechanisms for physical size and numerical value. In the ERP experiment, the cognitive load was manipulated in order to probe the degree to which information processing is shared across cognitive systems. As in the fMRI experiment, we found that the stimulus-stimulus conflict between numerical and physical magnitude is not completely resolved until response initiation. However, such late interaction was found only in the low cognitive load condition. In contrast, in the high load condition, physical and numerical dimensions interacted only at the comparison stage. We concluded that the processing of magnitude can be subserved by shared or distinct neural substrates, depending on task requirements.

Inhibition of return (IOR) refers to a reflexive mechanism mediated by phylogenetically primitive extrageniculate visuomotor pathways, which apparently serves to favor novel spatial locations by inhibiting those recently sampled. We demonstrate an asymmetry between temporal and nasal hemifields in the strategic modulation of IOR by endogenously controlled attention. Exogenous and endogenous precues were manipulated independently on each trial such that precues to initiate endogenous spatial orienting were presented after IOR had been activated by exogenous visual signals. Both types of precues manifested their characteristic effects on reaction time (RT) to detect subsequent targets: facilitation by endogenous precues, and IOR by exogenous precues. Under monocular viewing, an asymmetric interaction between these two mechanisms was observed. While endogenous allocation of attention to the nasal hemifield reduced IOR, no endogenous modulation of IOR was present in the temporal hemifield where the effects of the two types of precues were independent. These observations suggest a framework for understanding the neurobiology of automaticity and control—from an evolutionary perspective.