The robustness of the generalized signatures is also illustrated by our finding that, in signature extension, the number of days between the dates of acquisition of images used for training and the dates of images used for classification did not affect accuracy, as was also found in [2].
We observed very small technical variation within the same day, between different days of acquisition and between different acquisition sites.
Twenty-four hours after the last day of cued training, mice performed 5 days of acquisition training.
Across the 5 days of acquisition the animals, irrespective of genotype and age, showed a significant decrease in the escape latencies (Fig. 6a c) and the path lengths (data not shown) to reach the hidden platform (two-way repeated measures ANOVA, main effect of days: Escape latency: F (4,320) = 26.920; p < 0.001; Path length: F (4,320) = 28.148; p < 0.001).
Analysis of variance revealed that neither day of acquisition nor contrast agent used influenced the number of identified collateral arteries (p = 0.30 and p = 0.14, respectively).
Accuracy was always lower with signature extension compared to conventional classification without signature extension, but, for same season comparisons, we did not find an association between accuracy and the number of days between acquisition dates of images used in training and in classification (Figure 6).
Haemophilus ducreyi, a Gram-negative coccobacillus, is a strict human pathogen responsible for the development of chancroid, which is a sexually transmitted infection (STI) that causes the formation of genital ulcers 4 10 days after acquisition of the bacteria and persists in some cases for 1 3 months.
Mice were tested for 25 min each day over 12 consecutive days of training (Acquisition), and the number of correct and incorrect responses were recorded.
