Musical memory is the ability to recall music-related information, such as melodies and progressions of tones or pitches. Researchers have noted differences between linguistic and musical memory, leading to the theory that musical memory may be encoded differently from language and could represent an independent component of the phonological loop. However, this term's usage is problematic because it implies verbal input, whereas music is essentially nonverbal.

Consistent with hemispheric lateralization, there is evidence to suggest that the left and right hemispheres of the brain are responsible for different components of musical memory. By studying the learning curves of patients who have had damage to either their left or right medial temporal lobes, Wilson & Saling (2008) found hemispheric differences in the contributions of the left and right medial temporal lobes in melodic memory. Ayotte, Peretz, Rousseau, Bard & Bojanowski (2000) found that those patients who had their left middle cerebral artery cut in response to an aneurysm suffered greater impairments when performing tasks of musical long-term memory, than those patients who had their right middle cerebral artery cut. Thus, they concluded that the left hemisphere is mainly important for musical representation in long-term memory, whereas the right is needed primarily to mediate access to this memory. Sampson and Zatorre (1991) studied patients with severe epilepsy who underwent surgery for relief as well as control subjects. They found deficits in memory recognition for text regardless of whether it was sung or spoken after a left, but not right temporal lobectomy. However, melody recognition when a tune was sung with new words (as compared to encoding) was impaired after either right or left temporal lobectomy. Finally, after a right but not left temporal lobectomy, impairments of melody recognition occurred in the absence of lyrics. This suggests dual memory codes for musical memory, with the verbal code utilizing the left temporal lobe structures and the melodic relying on the encoding involved.

Platel (2005) defined musical semantic memory as memory for pieces without memory for the temporal or spatial elements; and musical episodic memory as memory for pieces and the context in which they were learned. It was found that two distinct patterns of neural activations existed when comparing semantic and episodic components of musical memory. Controlling for processes of early auditory analysis, working memory and mental imagery, Platel found that retrieval of semantic musical memory involved activation in the right inferior and middle frontal gyri, the superior and inferior right temporal gyri, the right anterior cingulate gyrus and parietal lobe region. There was also some activation in the middle and inferior frontal gyri in the left hemisphere. Retrieval of episodic musical memory, which includes music-evoked autobiographical memory, resulted in activation bilaterally in the middle and superior frontal gyri and the precuneus. Although bilateral activation was found there was dominance in the right hemisphere. This research suggests independence of episodic and semantic musical memory. The Levitin effect demonstrates accurate semantic memory for musical pitch and tempo among listeners, even without musical training, and without episodic memory of the original learning context.

Gaab, Keenan & Schlaug (2003) found a difference between males and females in the processing and subsequent memory for pitch using fMRI. More specifically, males showed more lateralized activity in the anterior and posterior perisylvin regions with greater activation in the left. Males also had more cerebellar activation than females did. However, females showed more posterior cingulate and retrosplenial cortex activation than did males. Nevertheless, it was demonstrated that the behavioural performance did not differ between males and females.

It has been found by Deutsch that lefthanders with mixed hand preference outperform righthanders in tests of short-term memory for pitch. This may be due to more storage of information on both sides of the brain by the mixed lefthanded group.

Experts have tremendous experience through practice and education in a particular field. Musical experts use some of the same strategies as do many experts in fields that require large amounts of memorization: chunking, organization and practice. For example, musical experts may organize notes into scales or create a hierarchical retrieval scheme to facilitate retrieval from long-term memory. In a case study on an expert pianist, researchers Chaffin & Imreh (2002) found that a retrieval scheme was developed to guarantee that the music was recalled with ease. This expert used auditory and motor memory along with conceptual memory. Together the auditory and motor representations allow for automaticity during performance, whereas the conceptual memory is mainly used to mediate when the piece is getting off track. When studying concert soloists, Chaffin and Logan (2006) reiterate that a hierarchical organization exists in memory, but also take this a step further suggesting that they actually use a mental map of the piece allowing them to keep track of the progression of the piece. Chaffin and Logan (2006) also demonstrate that there are performance cues that monitor the automatic aspects of performance and adjust them accordingly. They distinguish between basic performance cues, interpretive performance cues, and expressive performance cues. Basic cues monitor technical features, interpretive cues monitor changes made in different aspects of the piece, and expressive cues monitor the feelings of the music. These cues are developed when experts pay attention to a particular aspect during practice.

A savant is described as a person with a low IQ but who has superior performance in one particular field. Sloboda, Hermelin and O'Connor (1985) discussed a patient, NP, who was able to memorize very complex musical pieces after hearing them three or four times. NP's performance exceeded that of experts with very high IQs. However, his performance on other memory tasks was average for a person with an IQ in his range. They used NP to suggest that high IQ is not needed for the skill of musical memorization and in fact, other factors must be influencing this performance. Miller (1987) also studied a 7-year-old child who was said to be a musical savant. This child had superior short-term memory for music that was found to be influenced by the attention given to the complexity of the music, the key signature, and repeated configurations within a string. Miller (1987) suggests that a savant's ability is due to encoding the information into already existing meaningful structures in long-term memory.

Ruthsatz & Detterman (2003) define a prodigy as a child (younger than 10) who is able to excel at "culturally relevant" tasks to an extent that even isn't seen often in professionals in the field. They describe a case of one particular boy who had already released two CDs (on which he sings in 2 different languages) and was able to play several instruments by the age of 6.