Response preferences for ‘‘what’’ and ‘‘where’’ in human non-primary auditory cortex Doug J.K. Barrett * and Deborah A. Hall MRC Institute of Hearing Research, University Park, Nottingham, NG7 2RD, UK Received 15 December 2005; revised 16 March 2006; accepted 20 March 2006 Available online 2 June 2006 Primate studies suggest the auditory cortex is organized in at least two anatomically and functionally separate pathways: a ventral pathway specializing in object recognition and a dorsal pathway specializing in object localization. The current experiment assesses the validity of this model in human listeners using fMRI to investigate the neural substrates of spatial and non-spatial temporal pattern information. Targets were differentiated from non-targets on the basis of two levels of pitch information (present vs. absent, fixed vs. varying) and two levels of spatial information (compact vs. diffuse sound source, fixed vs. varying location) in a factorial design. Analyses revealed spatially separate responses to spatial and non-spatial temporal information. The main activation associated with pitch occurred predominantly in Heschl’s gyrus (HG) and planum polare, while that associated with changing sound source location occurred posterior to HG, in planum temporale (PT). Activation common to both pitch and changing spatial location was located bilaterally in anterior PT. Apart from this small region of overlap, our data support the anatomical and functional segregation of Fwhat_ and Fwhere_ in human non-primary auditory cortex. Our results also highlight a distinction in the sensitivity of anterior and posterior fields of PT to non-spatial information and specify the type of spatial information that is coded within early areas of the spatial processing stream. D 2006 Elsevier Inc. All rights reserved. Keywords: Auditory; fMRI; Functional segregation; Pitch; Spatial location Introduction One of the most dominant models of the auditory system suggests that object and spatial information remain independent up to and beyond the auditory cortex. This model originates from primate anatomy where two major routes have been identified, projecting from anterior and posterior belt and parabelt auditory fields to separate regions of the prefrontal cortex (Romanski et al., 1999). Differential response characteristics of cortical neurons within these routes lead Rauschecker and colleagues to propose that the two projections constitute distinct processing streams. Neurons in the anterior auditory belt are primarily responsive to the spectrotemporal features of a sound that code object identity and are consistent with a Fwhat_ pathway for object recognition; while neurons in the posterior belt are more sensitive to the spatial properties of a sound, consistent with a Fwhere_ stream for object localization (Rauschecker et al., 1995; Rauschecker and Tian, 2000, Tian et al., 2001, Rauschecker and Tian, 2003; see also Kaas and Hackett, 1999; Malhotra et al., 2004). The dual route model has been widely applied to interpret human data where potential homologies have been proposed for the organization of the primary auditory cortex and surrounding fields (Hackett, 2003). Support for anterior – posterior segregation has emerged from human anatomical studies (Galaburda and Sanides, 1980; Rivier and Clarke, 1997; Tardif and Clarke, 2001), functional magnetic resonance imaging (fMRI) (Bushara et al., 1999; Alain et al., 2001; Maeder et al., 2001; Warren and Griffiths, 2003), electromagnetic recordings (Alain et al., 2001; Anourova et al., 2001), and neuropsychological cases (Clarke et al., 2000). One specific case in point is that of pitch sensitivity. Evidence supports a highly conservative organization for pitch sensitivity in antero- lateral auditory cortex across both primates (Bendor and Wang, 2005) and humans (Griffiths et al., 1998, Patterson et al., 2002; Penagos et al., 2004; Hall et al., 2005). Responses in this region are elicited by both spectral and temporal pitch cues indicating a general cortical mechanism that is sensitive to both types of information. Recent evidence suggests that sensitivity to changing pitch sequences might also characterize posterior auditory regions in anterolateral planum temporale (PT), which is separate from activity in posteromedial PT that is elicited by sound sequences containing changing spatial locations (Warren and Griffiths, 2003). Such data imply divergent processing of spatial and non-spatial information even within PT, a cortical region that has previously shown to play a dominant role in spatial analysis (e.g., Baumgart et al., 1999; Warren et al., 2002). The division of spatial and non-spatial computational resources in the cortex is by no means an exclusive dichotomy (Middle- brooks, 2002; Recanzone, 2002). In the visual system, numerous cross-connections have been described between the two visual 1053-8119/$ - see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2006.03.050 * Corresponding author. Fax: +44 (115) 951 8503. E-mail address: doug@ihr.mrc.ac.uk (D.J.K. Barrett). Available online on ScienceDirect (www.sciencedirect.com). www.elsevier.com/locate/ynimg NeuroImage 32 (2006) 968 – 977