Outline 5.4 anatomical Ambiguities
- Core Argument dieses Kapitels definieren
To-Do’s
- check wording
- check sources
5.4 Anatomical Ambiguities
Assigning ROIs to the correct streams was anything but straightforward, given divergent accounts in the literature - most notably between Rolls et al. (2023) and Glasser et al. (2016). The former use effective connectivity to trace directional information flow, capturing output projections regardless of a region’s primary functional identity. The latter, on the other hand, rely on myeloarchitecture, cortical thickness, RSFC and task-fMRI activation profiles to classify areas by their dominant functional role. These methods answer different questions. Effective connectivity reveals that A4 and A5 maintain output channels to dorsal motion areas, while Glasser’s parcellation shows their dominant activation is language-related. Our resting-state partial correlation serves as a methodological middle ground: by partialling out the competing seed region, it isolates which prefrontal hub preferentially governs each region, and this is the criterion on which we resolved stream assignments.
Areas A4 and A5. Both showed robust coupling with IFJa; only A4 showed slight connectivity with FEF. Glasser’s argument where A4 and A5 are both activated in LANGUAGE STORY contrast - which is connected to IFG (44, 45, 47l) - supports the membership to the ‘what’-stream. Since A5 is also inferior to A4, this leads A5 more to the ventral pathway, which reflects the connectivity pattern. Whereas A4 might rather function as a router for information flow to higher level areas in mainly the ‘what’-stream. Rolls’ argument of A4/A5 demonstrating EC with MT and MST does not conflict with our results, since this reflects an indirect route to FEF. This might reflect a divergence hub, distributing information to both ‘what’ and ‘where’-pathways.
PSL. The PSL shows an interesting pattern, where it couples with IFJa left-hemispherically and with FEF right-hemispherically (Section 4.4.2). Additionally, Dureux (2024) argues that PSL is not responsive to auditory stimuli at all, which might reflect PSL as an abstract, modality-independent level, not on acoustic features. This is further supported by Rolls et al. (2023), who show directional EC from STS regions into PSL. PSL might serve as a supramodal connector or a high-level convergence hub, left side to semantic identity via IFJa, while right side links to spatial attention via FEF. It is interesting that a region which does not respond to any acoustic stimuli couples strongly with both prefrontal attention hubs, which could imply that top-down connections do not require bottom-up auditory drive at PSL.
STV. In the partial correlation analysis, STV couples with both prefrontal seeds at comparable strength (Section 4.4.3), making a clear stream assignment on connectivity strength alone impossible. Rolls et al. (2023), however, classify STV within the ventral language network, and Glasser et al. (2016) further report activation in social cognition and theory-of-mind task contrasts — both consistent with a primary what-stream affiliation. Together, these findings position STV as a multimodal interface within the ‘what’-stream, accessed by both top-down attention hubs.
Broca’s Areas 44 and 45. We used the Broca seed validation (Section 4.5.2) to analyse the connectivity patterns of both regions, since Rolls et al. (2023) argue for a dissociation where 45 belongs to an auditory ventral pathway, while 44 might form a dorsal stream. The results reveal a clear functional dissociation: Area 45 anchors the semantic ‘what’-stream via strong temporal coupling, while Area 44 orients toward motor-articulatory circuits via 55b and AVI. The cross-hemispheric 55b coupling of both areas suggests partial overlap rather than a strict boundary. Importantly, Area 44 shows no meaningful mean connectivity to FEF, despite a significant z-score, placing it within the ‘what’-stream rather than a dorsal spatial-motor framework.