Abstract: Retrieval practice is a learning strategy that improves memory efficiency by actively retrieving information from memory, typically through testing or free recall. Some clinical populations often present with memory and other cognitive impairments. This paper provides the first systematic review of research on retrieval practice in clinical populations, covering studies published between January 2020 and July 2025, and includes 22 studies in total. The findings indicate that, among developmental disorder populations, the positive effects of retrieval practice are manifested in promoting vocabulary memory in children who are deaf and hard of hearing (DHH), children with developmental language disorder (DLD), and patients with developmental dyslexia (DD), and full-text recall in patients with attention deficit/hyperactivity disorder (ADHD). Among acquired and neurodegenerative disease populations, this strategy helps improve long-term episodic memory in persons with Alzheimer's disease (AD), lexical-semantic comprehension in people with aphasia (PWA), and memory performance for names in stroke survivors (SS), as well as enhancing new information encoding efficiency in patients with schizophrenia (SP) and memory retention in patients with early psychosis (EP). Additionally, retrieval practice also exerts certain facilitative effects on health information memory in persons with HIV and word memory in children with Down syndrome (DS).
However, in certain disorders, its effects may be moderated. Firstly, preliminary evidence from studies of AD and SP suggests that the clinical benefits of retrieval practice may be moderated by the core cognitive deficits inherent to the disorder. Specifically, when deficits are predominantly at the encoding stage, external support may be needed for the retrieval practice to improve memory efficiency, whereas when deficits lie primarily in retrieval and interference control, retrieval practice itself can serve as a relatively effective rehabilitation approach. Secondly, evidence from children with DLD and patients with ADHD tentatively indicates that material complexity and ecological validity may moderate intervention outcomes. When material complexity exceeds patients' cognitive load capacity, the additional cognitive costs of retrieval may outweigh its memory consolidation benefits; meanwhile, retrieval practice tends to be more facilitative in contexts that resemble everyday learning. Thirdly, preliminary findings from patients with ADHD and SP indicate that the effects of retrieval practice may be shaped by other therapeutic interventions. For instance, the cognitive readiness afforded by pharmacotherapy could be an important prerequisite for retrieval practice to take effect, while external physical interventions (e.g., tDCS) that optimize specific brain region functioning may provide a favorable neural environment for it.
Future studies could further validate the moderating factors discussed above and focus on the following aspects. Firstly, future studies should investigate the benefits of retrieval practice for memory deficits in comorbid populations. Priority could be given to clinically prevalent comorbidity patterns, such as schizophrenia comorbid with anxiety disorders. This would help verify the effectiveness of retrieval practice and clarify its impact on shared cognitive deficits like memory impairments. Comorbidity characteristics may also guide intervention optimization in two ways: dynamically adjusting retrieval intervals to address memory decline, and targeting overlapping cognitive impairments across comorbidities to improve clinical relevance. Secondly, further exploration of the indirect effects of retrieval practice in clinical populations, such as strengthening metacognitive monitoring and self-efficacy, could open new avenues for clinical cognitive intervention. Thirdly, the roles of individual difference variables in retrieval practice among clinical populations warrant investigation, including working memory capacity, fluid intelligence, and effort, thereby providing an empirical basis for developing personalized intervention protocols tailored to patients' cognitive profiles.
To advance the clinical translation of retrieval practice, this paper proposes the following recommendations. Firstly, intervention goals should be clearly defined and appropriate materials selected accordingly. For example, for lexical-semantic comprehension deficits in PWA, minimal semantic pairs can be used, with each pair consisting of two semantically close but distinct concepts and their corresponding pictures. Secondly, strategy education should be integrated into the intervention process to enhance adherence. Rather than simple task training, the design of retrieval practice interventions may incorporate a comprehensive guidance system covering explanation of principles (knowledge transfer), hands-on experience (skill training), and positive reinforcement (motivational regulation). Thirdly, a stepwise implementation strategy is recommended. In the introduction phase, retrieval practice can supplement patients' existing cognitive training. For instance, at the end of each therapy session, patients may be asked to recall core knowledge points or events of the day. In the expansion phase, the dosage and task difficulty can be gradually increased, for example, by extending the duration of retrieval practice and adopting more functional, personalized materials such as memorizing new names or daily medication schedules. In the generalization phase, the emphasis shifts to strategy transfer, as patients begin to independently apply retrieval practice in daily life, such as spontaneously recalling the day's events before sleep. Through this process, retrieval practice can be internalized as an autonomous, self-regulated learning tool, thereby achieving the goals of clinical intervention.

Key words: retrieval practice, clinical, memory, cognitive rehabilitation