Note on the invariance of response latency in shuttlebox avoidance learning