Robert B Zajonc has combined neuropsychological interests with a leaning towards experimental social psychology. In the 1960s he revived interest in the everyday observation that people or objects you come across on a regular basis can "grow on you". His experiments were dubbed “mere exposure” effects. The stimuli were at first completely new to his participants, but repeated a varying number of times. He confirmed that, as the number of previous exposures to a stimulus increases from zero, so reported liking of it also increases. This setup establishes that the repetition causes the liking. A large literature followed upon his 1968 report. Reviewing it more than thirty years later Zajonc claimed “mere exposure” is a robust effect "across cultures, across species and across diverse stimulus domains".

I would say that caution and even alarm at something new may, as nothing aversive happens, give way to a “settling of nerves” that is experienced as pleasurable. But with more and more repetitions the liking reaches a peak and then turns downward - perhaps through boredom. It is also possible that liking increases only until the subject has “learned” the stimulus - there have been some suggestions that the exposure effect is easier to demonstrate with complex stimuli (and see Social and Behavioral Sciences Documents 16).

Both names and snatches of music, though at first manifesting an "exposure effect", beyond a certain degree of familiarity show an overexposure effect of decreased liking. This research is correlational rather than experimental. Some work of my own confirmed the basic results. Current Psychology 6 was experimental  (using synthetic speech as the stimulus) and though Journal of Environmental Management 21 was correlational it used a previously unresearched stimulus - landscapes.  The advantages conferred by microprocessors offering "do-loops" for the mere exposure experiment have transformed the ease of investigation. Machines that do not tire or vary can carry on repetition far beyond the maximum frequency used by Zajonc.

I have also found that brain damage from stroke reduces the exposure effects. Furthermore, amnesia is likely to influence exposure effects (Neuropsychology 4). This holds out the hope of an indirect way of assessing forgetting without having to elicit it - which is aversive to anyone.