Outline 2.2 The Auditory Where (“Dorsal”) Stream
- 2.2.1 Debate
- Rauschecker vs Hickock&Poeppel (where vs how)
- beide modelle erklären
- 2.2.2 Anatomical Definition & Evidence
- Glasser area A4, A5, Parietal Cortex
- Rolls Group 3
- 7Am, 7PC und so anschauen und verbindung mit FEF
- PF, PFcm, PFop connections anschauen
- 2.2.3 Synthesis
- Defining the where-stream as spatial/motion stream (maybe motor?)
To-Do’s
- [x]
2.2 The Auditory Where (“Dorsal”) Stream
In contrast to the ventral ‘what’-stream, the dorsal pathway has no single, universally accepted functional definition. Depending on the theoretical framework, the dorsal stream has been characterised as a spatial localisation system (Rauschecker & Scott (2009)), a sensorimotor integration pathway for speech (Hickok & Poeppel (2007)), or a system for processing affective prosody (Frühholz (2015)). These accounts are not mutually exclusive — they reflect different levels of description of the same anatomical pathway. For the purposes of this thesis, the dorsal stream is treated as a spatial-motor system subserving both spatial orienting and audiomotor integration, with the FEF as its prefrontal top-down control hub.
2.2.1 Spatial Processing and Audio-Motor Integration
The spatial-motor characterisation of the dorsal stream focuses on converging evidence across species and methodologies. Early support came from non-human primate studies demonstrating posterior auditory cortex sensitivity to sound location (Rauschecker & Afsahi (2023)), complemented by neuroimaging showing right-lateralized temporo-parietal activation for spatial processing (Griffiths et al. (1998)) and the clinical observation that spatial neglect is more severe after right-hemisphere damage.
Ahveninen et al. (2006) extended this account by reporting a double dissociation between anterior ‘what’ and posterior ‘where’ processing, with dorsal stream responses emerging at 70–150ms post-stimulus onset. Rauschecker (2011) further noted a cross-species shift: whereas non-human primates recruit the dorsal stream primarily for spatial localisation, humans appear to repurpose it for sensorimotor integration in speech.
This sensorimotor dimension was elaborated by Hickok and Poeppel (2004, 2007), who proposed that the dorsal stream maps acoustic speech signals onto articulatory representations, projecting from the STG via the Sylvian parietal temporal area (Spt) to frontal regions including Broca’s area (pars opercularis, Area 44, and pars triangularis, Area 45). Rather than contradicting the spatial account, this suggests that the dorsal pathway supports a broader spatial-motor function — localizing sounds in space and coupling that information with motor action plans.
2.2.2 Anatomical Definition and Evidence
The auditory dorsal “where”-stream originates posterior to Herschl’s Gyrus in the planum temporale and STG (Ahveninen et al. (2006)), and extends through auditory association areas before reaching parietal and frontal regions. A4 and A5 have EC to motion-sensitive areas MT and MST, which connect to superior parietal regions (7AL, 7Am, 7PC), consistent with a dorsal “where”-stream (Rolls et al. (2023)).
The Frontal Eye Field (FEF), located in the caudal middle frontal gyrus ventral to the junction of the superior precentral sulcus (sPCS) and superior frontal sulcus (SFS), serves as a primary prefrontal node of the dorsal stream. The FEF contains a full topographic map of contralateral space, predominantly encodes spatial information (cited from Bedini (2023): Wang et al., 2015) and is a core node of the Dorsal Attention Network (DAN), mediating covert spatial attention, oculomotor control, and spatial working memory (Bedini & Baldauf (2021)). It is structurally connected via SLF1 and SLF2 to posterior parietal (LIPd) and temporoparietal cortices, embedding FEF anatomically in the “where”-stream.
The inferior parietal cortex (specifically PF, PFop and PFcm) corresponds to the gyrus of the classical IPL (Baker (2018)) and forms a critical relay station within the “where”-stream. Consistent with both Rauschecker & Scott (2009) and Hickok & Poeppel (2007)/Hickok & Poeppel (2004), the “where”-pathway projects from posterior superior temporal cortex (pST) through parietal cortex before reaching prefrontal regions. These inferior parietal areas function as higher-level coordinator between auditory and somatosensory areas such as OP1-4 and FOP1 (Glasser et al. (2016)). Additionally, posterior parietal areas may serve as a relay station to the premotor cortex (PMC).
Importantly, the IPL does not appear to be driven by acoustic stimuli (Rauschecker & Scott (2009)), rather the angular gyrus is involved in higher-order speech processing - along with clear prefrontal activation. This suggests, the inferior parietal cortex is involved in domain-general, language-related speech comprehension, rather than low-level auditory processing (Rauschecker & Scott (2009)).
Finally, superior parietal regions 7AM, 7PC, 7AL receive direct auditory input from A4 and PBelt and may support tracking moving objects in space (Rolls et al. (2023)).
2.2.3 FEF’s Role in Auditory Spatial Attention
Salmi (2009) demonstrated FEF and PMC activation during both top-down controlled and bottom-up triggered shifting of auditory spatial attention. In their study, participants had to selectively attend to simultaneous tone streams at the left or right, occasionally shifting attention either in response to a visual cue (top-down, cue-guided attention shifts, CAS) or to loudness-deviating tones in the ignored stream (bottom-up, LDTs). Top-down controlled shifts activated bilateral FEF/PMC alongside SPL, IPS, TPJ, and IFG/MFG. Bottom-up triggered shifts activated an overlapping network - vmSPL, left IFG/MFG, left FEF/PMC, and right TPJ (Salmi, 2009) - showing that FEF recruits for auditory orienting tasks.
Considering the behaviour, deviants in the to-be-ignored-stream extended the response times, while deviants in the to-be-attended stream shortened them. This indicates that auditory bottom-up attention is modulated by the current focus of spatial attention. Interestingly, the same group had previously shown that shifting spatial attention in audition and vision recruits overlapping IPS/SPL and FEF/PMC regions Salmi (2007).
These findings suggest that FEF not only acts as a modality-specific oculomotor controller but might also follow a modality-independent principle, which the present thesis examines directly.
2.2.4 The “Where”-Stream as a Spatial-Motor System
Taken together, the evidence reviewed in 2.2.1, 2.2.2 and 2.2.3 suggests that the auditory “where”-stream does not map on one functional label only, but rather has a spatial-processing (Rauschecker & Scott (2009)) and a sensori-motor-integration (Hickok & Poeppel (2007)) function and reflect different levels within the same dorsal pathway.
Spatial processing in the temporo-parietal cortex is predominantly right-lateralized, consistent with the well-documented asymmetry of spatial neglect in patients with right-hemispheric damage (Rauschecker & Scott (2009)). In contrast, Hickok & Poeppel’s dorsal language model that links area Spt to Broca’s area is strongly left-dominant (Hickok & Poeppel (2007)). This hemispheric dissociation suggests that the “where”-stream operates in parallel in both hemispheres with different functional emphases: right-lateralized spatial and motion processing, and left-lateralized sensorimotor integration.
Further findings support this account, with caudal auditory belt area receiving somatosensory input (Rauschecker & Scott (2009)) and the inferior parietal cortex - especially PF, PFcm, PFop - serving as an interface between auditory, somatosensory and motor signals.
A related question concerns the dorsal language pathway, Rolls et al. (2023) report that connections of PBelt, A4 and A5 with BA44 may form a language-related dorsal pathway. However, this is not in conflict with a spatial-motor property: The same stream seems to perform spatial processing and sensorimotor feedback in the right hemisphere and articulatory and phonological processing on the left.
For the purpose of this thesis, the “where”-stream is therefore treated as a spatial-motor system: primarily organised around spatial and motor processing, with a left-hemispheric component that extends into speech production. The FEF serves as a primary prefrontal node, embedded anatomically in the DAN and structurally connected to the posterior parietal regions.