Benedikt Grothe (LMU) gives a talk at the MCMP Colloquium (30 October, 2013) titled "To map or not to map: or, how to represent auditory space". Abstract: More than 20 years ago, Walter Heiligenberg stated in a famous review that “Wherever we find behavioral responses guided by continual modulations of a certain stimulus variable, we seem to find an ordered representation of this variable within neuronal maps.” (Annu Rev Neurosci, 1991). While this concept may seem trivial for example in the visual system, where space is mapped directly onto the retina, auditory space is not represented at the receptor surface, and it is thought to be represented instead by a computed map from binaural disparity cues. Indeed, the ordered representation of auditory space computed from binaural disparities via a labeled-line code as described beautifully by Konishi, Knudsen and colleagues in the barn owl’s tectum served perhaps the ultimate test case for cartesian maps as a general solution for sensory coding of relevant cues. However, for mammals only crude “maps” representing large spatial areas at the single neuron level are found in their analog the superior colliculus and thus present a conceptual dilemma. Recent evidence indicates a dynamic population code for binaural disparities in mammals, which is incompatible with the concept of a fixed map of auditory space. Moreover, our most recent results show that self-regulated adaptation in the binaural detector neurons leads to drastic misjudgments of spatial position of sounds under predictable circumstances (Stange et al., 2013, Nat Neurosci) indicating relative rather than absolute spatial coding. Hence, computed maps seem to be only one possible solution for representing sensory information in the brain.