The insula (aka insular cortex, “Island of Reil”) is the cerebral cortex’s lonely island. Pushed deep into the lateral Sylvian fissure by the expansion of parietal and temporal lobes in human brain, the insula can not be observed from the brain surface and it’s separated from the neighboring claustrum by white matter (extreme capsule). It is relatively long rostrocaudally and recent studies have suggested the possibility of as many as 13 distinct functional subdivisions (Uddin et al, 2017).

Horizontal section of human brain. Red arrow points to insular cortex. Image from http://www.thehumanbrain.info
In rodents, the insular cortex is in generally the same anatomical location, but exposed on the lateral brain surface. The insula lies at the transition between isocortex (6 layers) and allocortex (3 layers): bracketed ventrally by the piriform (olfactory) cortex and secondary somatosensory cortex dorsally. Traditionally, insula subregions in rodent were defined based on the gradual presence of layer 4 (granule layer): granular insula (has layer 4), dysgranular insula (kinda has layer 4), and agranular insula (no layer 4). Over decades, investigations in rodents would identify the insula as a major site for higher-order visceral and gustatory signal processing, particularly within the granular and dysgranular cortex. In the mouse Allen Reference Atlas, the granular and dysgranular insular cortex were renamed as ‘visceral cortex’ (VISC) and ‘gustatory cortex’ (GU), with only the agranular insula (AI; itself having anterior dorsal/ventral vs. posterior subdivisions) retaining the insula name.

Coronal mouse ARA section 53 (left) with corresponding Neurotrace Blue tissue section (right). The insula includes agranular, dygranular, and granular subregions distinguished by layer 4 (the granule cell layer). The posterior agranular insula (AIp), adjacent to 3-layered piriform cortex (PIR), lacks a clear layer 4. The dysgranular insula (aka gustatory cortex (GU)) has a layer 4 but its granule cells are not quite fully laminar compared to granular insula (aka visceral cortex (VISC)). Other structures noted are SSs (secondary somatosensory cortex), claustrum (CLA), dorsal endopiriform nucleus (EPd).
The functional role of the insula is still a bit mysterious, despite the growing amount of research on this structure. In the human imaging field, the insula is a structure that wears many hats: it is implicated in a variety of roles including addiction, multimodal sensory processing, motor control, homeostasis, emotions (i.e. disgust, pain, anger, fear), empathy, saliency, and interoceptive awareness. With an involvement in so many different brain functions, it is difficult to understand exactly how the insula fits in to the function of the brain as a whole system.
The idea that the insula is involved in interoception — the conscious feeling of your internal state — has been particularly popular because it combines the insula’s role within gustatory and visceral circuits with the general perception of the cortex’s role in consciousness. Multisynaptic viral tract-tracing has shown the insula is positioned at the top of the pre-autonomic nervous system network –brain circuits which produce top-down control on autonomic neurons in the spinal cord/brainstem. In addition, the insula is densely connected to other parts of cortex, providing a gateway through which many cortical regions could alter homeostasis in the body. As is often the case, understanding the neural connectivity of a region gives you a pretty good of what it is doing functionally.
However, the insula is a prime example of where better anatomical understanding is needed if we are to take things further. In both human and rodent research, the insula is a large structure with multiple uncharacterized subdivisions. As you can see from the picture above in the mouse, the VISC, GU, and AI subregions are particularly thin and difficult to target selectively via injection. Many behavioral lesion studies often destroy the entire insula (and more), leaving the resulting interpretation ambiguous as to which subregion is really involved. In the human imaging field, there is an effort to define subdivisions, but MRI is limited in resolution and often cannot localize effects to small areas. As tools become more specific and we can obtain higher resolution, the need for anatomy will become more important than ever. Those who are willing to appreciate the details of the anatomy will be those who make the next big break-through.
-M














