Background: Rizzolatti's group has identified that neurons in the inferior frontal cortex of monkeys are activated during the observation of a finalistic action performed by another monkey, as well as during the execution of the same action. In humans, similar activations have been described with fMRI during observation and imitation of fnalistic motor actions in the left inferior frontal cortex as well as in the rostral-most region of the right superior parietal lobule (Iacoboni et al). These areas have been described as the "mirror system", a system matching observation and execution of finalistic goal-related motor actions. Recent fMRI data on observation-imitation of visual emotional expressions have observed the activation of the insula and amygdala as well as of the areas of the mirror system (Carr et al, in preparation). These data suggest the existence of a direct-matching system involved in emotions internal representation, very similar to the mirror system for motor actions. In this perspective, insula seems to play a fundamental role as a relay between mirror areas and limbic circuitries. Anatomical lesions of different components of the mirror system and of its links to the limbic structures may therefore cause an impairment of emotional processing. In the present study we investigated whether strokes involving the insular region impair emotions recognition.

Methods: Insular stroke, non-insular stroke patients and controls were submitted to a version of the Ekman faces test, modified for italian people. The test consists of six different emotions (happiness, sadness, anger, surprise, disgust, afraidness) to be recognized in 24 photos presentation, either "faces", "eyes" or "mouths".

We studied 13 patients with single ischemic strokes involving the insular region (IP) (age 65,2±10,4; M/F: 5/8; education 6±2,3; USS 26,4±3,2), 13 "non-insular" single stroke patients (NIP) (age 61,6±9; M/F: 6/7; education 7,5±3,5; USS 25,8±5,6) and 17 normal controls (age 62±4; M/F: 8/9; education 6,2±2). All patients were non-demented, underwent a neuropsychological work-up to exclude agnosic or attentive deficits and were submitted to the Benton test, in order to exclude a selective disorder in unknown faces identification.

Results: A significant impairment in visual emotion recognition at the Ekman test was observed in all cerebrovascular patients (IP+NIP), regardless of the location of the single lesion compared with controls (Ctrl: 18,5±2,2 vs CVD: 13±4; p=.0001). Furthermore, IP presented a more severe impairment (10,6±4) in emotions recognition when compared both with NIP (15,4±1) as well as with controls (18,5±2,2)(p<.0001).

Conclusions: Stroke patients show an impairment in visual emotions recognition, regardless the site and side of the lesion. However, when the lesion affects the insula, the visual emotions recognition is drammatically lost in comparison with patients with lesions affection other cerebral regions. These preliminary data support the role of the insula as a relay in the pathways subserving emphatic resonance, and indicate the relevance of "mirror system" not only for motor actions, but even for emotional behaviour.