BackgroundNeurocognitive impairments are widespread in individuals with schizophrenia both premorbid and post illness onset. Although less pronounced, individuals with bipolar disorder also display various neurocognitive deficits. Owing to the impaired neurocognitive functions in first-degree relatives, neurocognitive deficits are considered endophenotypes of both disorders. Importantly, neurocognitive heterogeneity exists in both disorders. Distinct neurocognitive subgroups in young adults with clinical high risk of psychosis or familial risk of bipolar disorder have been identified with an increased risk of conversion to illness and poorer functioning in the most neurocognitively impaired subgroup. However, neurocognitive heterogeneity remains to be investigated in young children at familial high risk of schizophrenia (FHR-SZ) and bipolar disorder (FHR-BP). The aim of this study was to identify relatively homogeneous neurocognitive subgroups in a cohort of children with FHR-SZ or FHR-BP and to investigate the distribution of high risk status across the neurocognitive subgroups. Exploratively, potential differences between these subgroups on functional outcome and polygenic risk were investigated.MethodsNeurocognition was assessed cross-sectionally with a comprehensive test battery in a population-based cohort of 514 children aged 7 with either FHR-SZ (N=197), FHR-BP (N=118) or no familial high risk of these disorders (N=199). A hierarchical cluster analysis of the neurocognitive data was performed using Ward’s method and followed-up by a K-means cluster analysis. The neurocognitive clusters were further compared on measures of current level of functioning and polygenic risk for educational attainment.ResultsThree distinct neurocognitive subgroups were derived from the hierarchical cluster analyses: a 1) significantly impaired, 2) typical, and 3) high functioning subgroup. The three subgroups performed significantly different from each other on the majority of the pairwise comparisons of the neurocognitive functions (Cohen’s d range = 0.33–2.27, p < 0.01). There was a significantly higher percentage of children with FHR-SZ in the significantly impaired subgroup (36%) compared to children with FHR-BP (20%). Importantly, 64% of the children at FHR-SZ and 80% of the children at FHR-BP displayed either typical or high neurocognitive functioning. Finally, the neurocognitive subgroups differed significantly on current level of functioning with the significantly impaired subgroup displaying the lowest level of functioning. Polygenic risk for educational attainment was non-significantly different across neurocognitive subgroups.DiscussionYoung children at familial high risk of severe mental disorders (SMD) are neurocognitively heterogenic with three distinct neurocognitive subgroups of varying severity. A higher percentage of children at FHR-SZ was in the significantly impaired subgroup (36%) compared to children at FHR-BP (20%). In this cross-sectional study, being in the significantly impaired neurocognitive subgroup is associated with poorer daily functioning across risk status and in other, longitudinal studies has proven predictive of transition to illness in young adults at clinical high risk. Therefore, neurocognitive profiling in high-risk offspring may guide future intervention programs targeting the neurocognitively impaired subgroup of young children at familial high risk of SMD using e.g. cognitive remediation or other potentially preemptive or ameliorating interventions.