The Insula and Evaluative Processes

Study at a glance

• Authors: Gary G. Berntson, Greg J. Norman, Antoine Bechara, Joel Bruss, Daniel Tranel, John T. Cacioppo
• Year: 2010
• Journal: Psychological Science
• DOI: 10.1177/0956797610391097
• N: 28
• Design: lesion / neuropsychological case series (three groups: insula, amygdala, control lesions)

Key findings

• Insula-lesion patients (n=7) showed attenuated arousal ratings to affective pictures — most apparent at the pleasant and unpleasant extremes, neutral spared — for both hedonic signs, and attenuated valence ratings (smaller increments in positivity as pleasant stimuli grew more pleasant, and in negativity as unpleasant stimuli grew more unpleasant).
• Amygdala-lesion patients (n=12) showed the opposite, narrower profile: valence ratings indistinguishable from controls, with arousal selectively attenuated for unpleasant stimuli only (the typical arousal increment for pleasant pictures was preserved).
• A double dissociation across both dimensions of evaluation: the insula plays a broad role spanning valence and arousal for both hedonic signs; the amygdala a selective role in arousal, especially to aversive stimuli.
• Insula patients’ neutral-stimulus ratings, and their self-ratings on the UCLA Loneliness and CES-D scales, were normal — so the deficit is not a general under-rating bias but a failure to register/categorize the affective content itself.

Limitations

• Very small, heterogeneous groups (7 insula, 9 control, 12 amygdala).
• The amygdala group is confounded with the others: different aetiology (anterior temporal lobectomy for seizures vs. ischaemic stroke), significantly younger (age covaried), and higher depression scores — hence treated as ancillary.
• Insula lesions were large MCA-territory strokes involving surrounding tissue, unilateral, and split L/R (4R, 3L) — too few to test the laterality Craig’s asymmetry hypothesis would predict.
• Evaluations are self-report on rating grids; arousal was rated after the picture was removed (memory deficit ruled unlikely given intact standard memory scores). Whether the intended valence/arousal dimensions were directly measured is uncertain.

Connections

• Methods: Affective picture viewing, Lesion-symptom mapping
• Concepts: Insular cortex, Amygdala, Core affect
• Debates: Do discrete emotions map onto distinct brain regions or networks?
• Related studies: Jones et al. (2010), Lindquist et al. (2012), Haruki & Ogawa (2023), Bechara & Damasio (2005), Dunn et al. (2010)

A rare thing on this wiki: a causal, lesion-based dissociation of the two dimensions of core affect — valence and arousal — across two structures. Where almost all the wiki’s valence/arousal evidence is correlational (cardiac coupling, heartbeat-accuracy group differences, lexical norms), Berntson and colleagues remove the structures and watch which dimension goes. See Lesion-symptom mapping for what that design can and cannot license.

The design

Twenty-eight lesion patients from the University of Iowa registry viewed IAPS pictures spanning five categories (very/moderately unpleasant, neutral, moderately/very pleasant) and rated each on positivity and negativity (a bivariate 5×5 grid) and on arousal (a 7-point scale), using the affective picture paradigm with a bivariate valence grid rather than the usual unidimensional SAM. Three groups:

• Insula (n=7): unilateral ischaemic stroke destroying >50% of the insula (anterior and posterior) — the primary group.
• Control (n=9): strokes sparing the insula, amygdala and other affect-critical regions — the primary contrast.
• Amygdala (n=12): anterior temporal lobectomy for seizure control — an ancillary contrast, confounded by aetiology and age.

The double dissociation

Insula lesions blunt evaluation broadly. Compared with control-lesion patients, insula patients gave attenuated arousal ratings to affective pictures — the deficit sharpest at the very-pleasant and very-unpleasant extremes (F(1,14)=6.99, d=0.91), with normal ratings to neutral pictures. Their valence ratings were also attenuated: smaller increments in positivity as pleasant stimuli became more pleasant, and in negativity as unpleasant became more unpleasant (significant quadratic-trend group differences on both). The loss spans both dimensions and both hedonic signs.

Amygdala lesions blunt one corner. Amygdala patients’ positivity and negativity ratings were statistically indistinguishable from controls (d < 0.01). Their arousal deficit was selective for unpleasant stimuli — a typical arousal increment survived for pleasant pictures, but not for increasingly unpleasant ones — replicating the authors’ earlier finding in a differently-lesioned amygdala sample (Berntson et al. 2007).

So: the insula is required broadly, for valence and arousal across positive and negative content; the amygdala is required selectively, for arousal to aversive content, while leaving the recognition and categorization of valence intact.

Where it lands in the wiki

On the Insular cortex page. This is direct, causal evidence that the insula is “broadly involved in the recognition, processing, and assignment of evaluative valence, and may contribute to affective arousal” — the core-affect role, arrived at by subtraction rather than activation. It sits beside Jones et al. (2010) as the wiki’s second dedicated insula-lesion source, and complements the imaging deflations (Haruki & Ogawa) that the (right) anterior insula is about arousal/affect more than the body in general: here, removing the insula costs both affective dimensions, so its evaluative role is not confined to arousal. Note the usual caveat — large MCA strokes, small n, no laterality test.

On the Amygdala page. A clean instance of the amygdala’s selective role in negative arousal — consistent with reading (2) (salience/arousal detector, arousal-preferring for aversive content; Adolphs, Russell & Tranel 1999) and with reading (4)‘s emphasis on the amygdala coupling stimuli to arousal. Critically, it dissociates the amygdala from valence: these patients recognized and categorized the negative content (normal valence ratings) yet under-rated its arousal — “registering the arousal or emotional impact of stimuli” without determining whether a stimulus is appetitive or aversive.

On Core affect. A causal complement to the four-source correlational table. Those studies show felt arousal is the bodily/interoceptive dimension while valence has a separate provenance. Here the two dimensions come apart again, but anatomically: the amygdala touches only arousal (and only its negative pole), sparing valence — a structural echo of “arousal is the bodily dimension, valence is a further story.” The insula, tellingly, touches both, which fits its position as the integrative apex rather than a pure arousal relay.

A note on the researcher

This is Berntson’s empirical, autonomic-psychophysiology lineage showing through — the same author whose 2021 anatomical review the wiki holds. The 2010 study is co-authored with Antoine Bechara, Daniel Tranel and the Iowa lesion group, tying the insula/amygdala evaluative work to the same patient registry and method that produced the somatic-marker literature (Bechara & Damasio (2005), Bechara et al. (2000)).