Fig. 1

Change of Mini-Mental State Examination (MMSE) score over time. Patients were grouped by whether they were prescribed neural network treatment of choice (NNToC) according to the Long Short Term Memory (LSTM) model. The X-axis represents the duration of taking a particular medication, where x = 0 means the treatment initiation time. Data shown are mean values, with error bars indicating standard deviation. A Internal validation on Southern Health NHS Foundation Trust (SHFT). Two hundred eighty-five [21.22%] patients were prescribed NNToC and reported a smaller MMSE reduction after 2 years compared to the 1058 [78.78%] patients who were not (0.60 [0.26] vs 2.80 [0.28], respectively; P = 0.02). B External validation on Oxford Health NHS Foundation Trust (OHFT). Two hundred twenty-two [12.53%] patients were prescribed NNToC and reported a smaller MMSE reduction after 2 years compared to the 1550 [87.47%] patients who were not (1.01 [0.49] vs 4.23 [0.60], respectively; P = 0.01). A detailed quantitative report of score changes is provided in Additional file 1: Table S5