Fig. 9 | Nature Communications

Fig. 9

From: Deviation from the matching law reflects an optimal strategy involving learning over multiple timescales

Fig. 9

Data shows that the LMIT is affected by the length of recent consecutive experimental sessions (model-independent analysis). a, c LMIT increases as the monkeys completed more trials in recent days. The LMIT and the mean recent experimental day length (how many trials monkeys completed in one experimental day) show a significant positive correlation [permutation test: p < 0.025 for Monkey F and p < 0.004 for Monkey G]. b, d LMIT decreases as monkeys had longer breaks between experimental sessions. The LMIT and the mean recent break length show a significant negative correlation [permutation test: p < 10−6 for Monkey F and p < 10−5 for Monkey G]. The LMIT is computed in the same way as the previous figures with sliding windows, and both the mean recent experimental day length and the mean recent break length are computed as the average over the past 12 experimental days. Please see Supplementary Fig. 15 for the time course of changes in these variables, and also Supplementary Fig. 16 for the correlations with wSlow, instead of LMIT

Back to article page