Comparing exponential race and signal detection models of encoding stimuli into visual short-term memory.