The position of the red square changed across trials, rotating clockwise, and indicated the position of the trial in the sequence of alternating runs of two trials.
Each recording session consisted of two 40-min trials, with the compartments' positions switched and the floor of the environment rotated between trials to control unidentified sensory cues distinguishing the compartments.
On each-training day the animals swam four trials (rotating initial placement each time), lasting one minute each, with a five minute interval between trials.
The average latency to fall from the rotating rod in the test trials (average of the nine runs in the three trials) was calculated for each mouse.
Performance on the Rota-Rod was measured by latencies to falling off the rotating cylinder, during 10 trials over 1 day, with a 5-min inter-trial interval.
The videos of the rotating box preceding the trials were considered baseline and not modeled separately.
In addition, the pool was rotated after every trial (Horne et al., 2012; Pearce et al., 2004).
As the cross-maze was rotated between every trial it is most unlikely that intra-maze cues could have assisted.
The maze was rotated after each trial in a random sequence of 90°, 180° and 270° to remove intra-maze cues.
To confirm that similar distribution asymmetries occur with a non-biological stimulus, we ran three observers on trials with the rotating walker and an ambiguous rotating cuboid.
Instead, eleven subjects solved such trials by mentally rotating the horizontal circle pair between encoding and retrieval with a 90° angle.
