Error patterns on a computerized version of Raven's progressive matrices in specific learning disorders.

Specific Learning Disorders (SLDs) are neurodevelopmental conditions characterized by persistent difficulties in reading, writing, or mathematics. These difficulties occur in the absence of intellectual disabilities, sensory impairments (visual/auditory), or neurological issues, indicating that SLDs do not reflect a lack of ability or a global cognitive impairment. Rather, they represent a different learning profile, where domain-specific difficulties coexist with otherwise adequate intellectual functioning. Consistent with this perspective and contemporary models conceptualizing intelligence as a multidimensional construct, research reveals marked heterogeneity in SLD cognitive profiles, particularly subtle weaknesses in working memory and processing speed, closely intertwined with Fluid Intelligence (FI). FI is defined as the ability to think logically, identify patterns, and solve novel problems independently of prior knowledge. Existing research has not found different FI levels in SLDs; however, most studies focused exclusively on quantitative measures, such as total accuracy, which may hinder individual differences in cognitive processes. Qualitative dimensions of FI, particularly error patterns, remain largely unexplored in this population, yet examining these patterns could reveal inefficiencies in reasoning that quantitative scores cannot capture. The present study investigated FI in children and adolescents with SLD using MatriKS, a new computerized version of Raven's Progressive Matrices. A sample of 160 participants aged 7-19 years (88 with SLD, 72 typically developing) completed MatriKS alongside standardized cognitive and academic assessments. Among participants with SLD, further distinction was considered, based on academic impairment area (reading/writing disorder vs. mathematical disorder, isolated or combined). Results demonstrated good convergent validity of MatriKS with traditional Raven's Matrices, and revealed significant group differences in overall accuracy, with the mathematics subgroup showing the most pronounced difficulties. Qualitative error analysis uncovered distinct patterns across SLD subtypes and developmental stages, with specific error types successfully distinguishing between diagnostic groups at different ages. These findings show that while global FI scores may fall within average range, qualitative error analysis can uncover specific inefficiencies, differentiating SLD subtypes across development. Therefore, incorporating error analysis into cognitive assessment could enhance the understanding of individual cognitive profiles and provide insights for interventions addressing underlying processing vulnerabilities, rather than solely academic skill deficits.