Topography of claustrum and insula projections to medial prefrontal and anterior cingulate cortices of the common marmoset (Callithrix jacchus)
We recently hypothesized that the claustrum may have an active role in modulation or switching between resting‐state functional networks (Reser et al., 2014). These networks are defined largely by synchronous changes in blood flow (Fox and Raichle, 2007; Raichle, 2015) or oscillations in field potentials (Srinivasan et al., 2007; Jerbi et al., 2010). Areas that form the main hubs of these networks are known to be affected in a variety of psychiatric and neurological conditions (Crossley et al., 2014; Stam, 2014). For the present study we placed neuroanatomical tracer injections in cortical areas along the cerebral midline of the common marmoset, from the medial prefrontal to the caudal anterior cingulate areas. These regions are of particular interest because they include hubs of the default mode and salience resting‐state functional networks (Seeley et al., 2007; Raichle, 2015).
A key issue that we have addressed is the anatomical relationship between the claustrum, the adjacent anterior insular cortex, and the component areas of the salience resting‐state network. Neuroimaging studies have indicated that the anterior insular cortex is one of the main hubs of the salience network, implicated in the transition from task‐negative to task‐positive networks (Menon and Uddin, 2010). However, studies in rodents have demonstrated strong connections between the claustrum and cingulate and medial prefrontal areas (Sloniewski et al., 1986; Hoover and Vertes, 2007; Mathur et al., 2009), including those that could form a homolog of the primate salience resting‐state network (Goll et al., 2015; Belcher et al., 2016). Given the proximity between claustrum and insular cortex and the relatively low resolution of the noninvasive imaging methods that are directly applicable to the human brain, it is uncertain whether the claustrum, anterior insular cortex, or both of these structures originate inputs to salience network areas. The claustrum and insular cortex are not only closely apposed (Kapakin, 2011) but also share a common arterial blood supply (Marinkovic and Markovic, 1990; Ture et al., 2000; Delion and Mercier, 2014). The venous drainage of the claustrum is also shared with surrounding structures, including the internal, external, and extreme capsules; frontoparietal white matter; and striatum (Zhang et al., 2015), potentially confounding assignment of functional activity as measured by the BOLD signal (Boubela et al., 2015). Furthermore, the claustrum is prone to partial volume artefacts in functional imaging studies (Wu et al., 2010; Hornak, 2014) because it is quite thin in cross‐sectional area (Kowiański et al., 1999; Druga, 2014) and is situated between two white matter tracts. It is thus critical to examine its connectivity with higher resolution methods, in the context of the proposed functional networks.
The marmoset is an ideal model species for this study; it is a small primate that lacks a cingulate sulcus, which simplifies injection placement.