BackgroundPhosphatidylserine (PS) and α-Linolenic acid (ALA), are positively associated with cognitive function, but their combination effects and possible mechanisms remain unclear. We aimed to explore the effects on cognition and potential mechanism of the supplements. MethodsThis randomized, double-blind, placebo-controlled trial recruited 190 MCI patients in Tianjin, China, and randomly assigned in intervention group and placebo group. Each group consumed two capsules every day for 12 months. Each capsule for intervention group contains 144 mg ALA, 31.5 mg PS and 3.6 mg Ginkgo total flavonoids as main functional components, with 0.48 mg Vitamin B1 (as thiamine hydrochloride), 0.48 mg Vitamin B6 (as pyridoxine hydrochloride) and 90 μg folic acid as supplement. Capsules for placebo group were identical but contain no active ingredients. Cognitive function, serum n-3 polyunsaturated fatty acids (PUFAs) and neurotransmitters were assessed at baseline and 12 months. Linear mixed effects model and causal mediation analysis were conducted to explore the effects and potential mechanism of the intervention. ResultsA total of 190 participants (mean [SD] age, 67.95 [5.62] years; 70 (36.8 %) male and 120 (63.2 %) female) were randomized to the placebo group (n = 95) and intervention group (n = 95). Compared with placebo group, the intervention group had statistically significant improvements in arithmetic testing (β, 0.688; 95 % CI, 0.103–1.274), the similarity test (β, 1.070; 95 % CI, 0.472–1.667) and short-term memory (β, 0.600; 95 % CI, 0.399–0.800). Besides, the intervention group had statistically significant increases in serum ALA (β, 1.620; 95 % CI, 0.967–2.265), DHA (β, 2.797; 95 % CI, 1.075–4.532), EPA (β, 1.472; 95 % CI, 0.296–2.643), acetylcholine (β, 0.441; 95 % CI, 0.415–0.468), GABA (β, 0.009; 95 % CI, 0.001–0.016) and 5-HT (β, 0.160; 95 % CI, 0.081–0.238) compared to the placebo group. And the intervention may improve short-term memory by increasing serum ALA levels (average causal mediation effect = 0.132, 95 % CI, 0.053–0.225) with 19.7 % mediation proportion. ConclusionsThis food supplement containing phosphatidylserine could improve different cognitive functions of MCI patients, especially short-term memory, and increase serum n-3 PUFAs and neurotransmitters levels. Serum ALA level might play a mediation role.