Table 4.

Effect of behavioral parameters on the results shown in Fig. 7C

Independent Variable 1	Independent Variable 2	Dependent Variable	r2 (1)	p (1)	m (1)	p (2)	m (2)
Mean N within session	N/A	Mean N between session	0.93	0.005	1.66	N/A	N/A
Mean N within session	Mean speed difference N between session	Mean N between session	0.97	0.04	1.22	0.15	−0.12
Mean N within session	Mean acceleration difference N between session	Mean N between session	0.95	0.05	1.32	0.30	−0.06
Mean N within session	Mean absolute angular velocity difference N between session	Mean N between session	0.91	0.04	1.76	0.75	0.0013
Mean N within session	Mean speed N between session	Mean N between session	0.92	0.05	1.49	0.57	0.01
Mean N within session	Mean acceleration N between session	Mean N between session	0.93	0.05	1.47	0.49	0.01
Mean N within session	Mean absolute angular velocity N between session	Mean N between session	0.94	0.02	1.66	0.35	−0.0018

We constructed a total of 6 new GLMs (rows 2-7) to control for the findings shown in Fig. 7C (row 1). Each time, we recreated the GLM of Fig. 7C but added a second independent variable to the model to examine whether the behavior of the animals could explain the structured heterogeneity in their partial remapping. Across all 6 controls, the correlation between the two types of neural data (independent variable 1 and dependent variable) was high, as indicated by the non-zero slope values (m1). The relationship between the neural variables remained significant (p(1) < 0.05) even after the incorporation of the behavioral controls. However, we found no correlation between the behavioral data and the neural data (independent variable 2 and dependent variable), as indicated by the negligible slope values (m2). The relationship between the behavioral and neural data was not significant (p(2) > 0.05), confirming that the variability in the animals' remapping characteristics cannot be explained by the variability in these behavioral metrics.