People with Korsakoff's syndrome perform substantially worse than control participants at recognising emotion in faces and in tone of voice, according to the first meta-analysis to pool findings across the levels of social cognition in the condition. Kyra Wijnen and colleagues at the Vincent van Gogh Institute for Psychiatry in Venray and the Donders Institute at Radboud University, writing in Addiction (published online 18 November 2025, doi 10.1111/add.70256), report a large weighted effect for emotion perception (Hedges' g = -1.14, 95 per cent confidence interval -1.46 to -0.81, drawn from 8 studies covering 372 participants), with comparable effects for facial emotion recognition and for prosody, the emotional colouring of speech. The literature search was carried out in May 2024, and the comparison in the pooled studies was between people with alcohol-related Korsakoff's syndrome and controls who had neither cognitive impairment nor alcohol use disorder.

The picture beyond emotion perception is more mixed. Mentalising, the work of inferring what another person is thinking or intending, showed a large effect (g = -1.05, 95 per cent CI -1.61 to -0.50) but rested on only 3 studies and 120 participants. Social knowledge, meaning grasp of social rules and conventions, showed a large effect (g = -0.81, 95 per cent CI -1.24 to -0.38) from a single study of 104 people. By contrast, the results for empathy (g = -0.43, 95 per cent CI -1.05 to 0.20, from 1 study of 40 people), self-awareness (g = -0.21, 95 per cent CI -0.47 to 0.05, from 1 study of 68 people) and moral reasoning (g = -0.54, 95 per cent CI -1.28 to 0.19, from 2 studies of 80 people) all crossed zero, which the authors describe as uncertain, with possible important differences in both directions.

Korsakoff's syndrome is a permanent, non-progressive amnestic disorder that follows thiamine deficiency, usually in the context of heavy alcohol use, as set out in a 2017 review by Arts, Walvoort and Kessels in Neuropsychiatric Disease and Treatment. It is dominated by memory loss, apathy, confabulation and poor insight, and it is commonly encountered by care staff on specialist wards and in long-stay units. For relatives and staff, the practical reading of the new figures is that a person who does not respond to a worried face or a warning tone may be failing to register the signal rather than choosing to ignore it, and that guidance on communication has firmer evidence behind it for emotion recognition than for empathy or insight.

The counterpoint is the size of the evidence base itself. Several of the sublevels rest on one or two studies, some pooled estimates were highly heterogeneous (96.6 per cent for social interpretation overall), and formal tests for publication bias could not be run for the smaller analyses because too few studies existed. The authors' own conclusion is limited to the statement that performance in emotion recognition, mentalising and social knowledge is poorer in people with alcohol-related Korsakoff's syndrome than in controls, with medium to large effect sizes.