Cognitive tasks for evoking the P300 event-related potential in Alzheimer's disease and mild cognitive impairment: A scoping review

Event-related potentials (ERPs), particularly the P300 component, provide a noninvasive, cost-effective approach to probing neural dysfunction in Alzheimer's disease (AD) and mild cognitive impairment (MCI). However, methodological and task-related variability across studies has hindered clinical translation. This scoping review systematically mapped cognitive tasks, stimulus parameters, and EEG acquisition pipelines used to elicit the P300 in AD and MCI and assessed their consistency in revealing group differences in P300 latency and amplitude relative to healthy controls. The review followed Joanna Briggs Institute methodology and PRISMA-ScR guidelines. Searches across PubMed, ProQuest, and ScienceDirect (2015-2025) identified empirical studies that employed EEG-based P300 paradigms to compare AD, MCI, and healthy older adults. Data on participant characteristics, task type, stimulus parameters, ERP outcomes, and EEG processing were extracted. Twenty-eight studies met the inclusion criteria. Three dominant paradigm groups were identified: auditory and visual oddball, working memory, and selective attention-executive tasks. Across studies, AD most consistently demonstrated prolonged P300 latency - most robustly in auditory oddball paradigms - whereas amplitude reductions were less reliable. In MCI, reduced amplitude was most consistently observed relative to controls under higher cognitive load, while prolonged latency emerged in selected visual oddball and selective attention-executive tasks. Stimulus duration (50-150 ms), moderate target probability (∼20-30%), behavioral response, and sufficient trial counts enhanced sensitivity. Methodological heterogeneity in referencing, filtering, artifact rejection, and component quantification (e.g., peak vs mean amplitude) limited reproducibility. The P300 remains a promising electrophysiological biomarker for early cognitive decline. Latency prolongation might reflect AD-related slowing of processing, while amplitude reduction under cognitive load may capture MCI-related inefficiency of resources. Standardization for task design and EEG processing will be critical for establishing P300 as a clinically useful marker of pre-dementia progression.