Skip to main content
Fig. 5 | Journal of Translational Medicine

Fig. 5

From: Realising stratified psychiatry using multidimensional signatures and trajectories

Fig. 5

a Trajectories represented as a graph with nodes shown for clusters A, B and C at baseline (time 1) and clusters 1–5 at follow-up (time 2); heavier line weights between clusters denote higher probability of transitioning to each cluster at time 2, given the patient was in one of clusters A, B or C at time 1 b Table of probabilities corresponding to line weights in a. c Bivariate plot (Time 1) reproduced from Column B of Fig. 4. Baseline patient signatures at time 1, with diamonds showing approximate projected location of prototypes of clusters; a single ‘case’ patient signature is represented by a black dot. d The probability distribution of the ‘case’ patient belonging to clusters A, B or C (in 5 dimensional space—note that projecting into 2 dimensions results in distortion which visually misrepresents the nearest prototype). e At time 2, the ‘case’ patient’s signature demonstrates a trajectory of improvement in both verbal memory and vigilance performance. f Probability distribution of the ‘case’ patient’s membership to the clusters at time 2 given (conditional on) the probability the test patient was in cluster B at time 1

Back to article page