Musical perception can predict learning difficulties in children

Researchers create tests to assess musical perception and use the results as a basis for identifying children predisposed to problems with written and spoken language acquisition.

By Karina Toledo  |  Agência FAPESP – A test that measures musical perception in children aged 6-13 years has been developed by a group of scientists from Brazil, Canada, the United States and the United Kingdom.

The test and other results of their research, which was supported by FAPESP, were published in January in the journal Frontiers in Neuroscience.

“Our goal in the future is to demonstrate that the same test can identify children predisposed to difficulties with speech and writing, which would help parents and teachers intervene early on,” said Hugo Cogo Moreira, a professor and supervisor in the psychiatry and medical psychology graduate studies program at the Federal University of São Paulo (UNIFESP) in Brazil and the principal investigator for the project.

According to Moreira, there is evidence in the scientific literature to support the hypothesis that musical perception in children – their ability to perceive such properties of music as pitch, loudness, melody and rhythm – correlates with verbal language learning skills.

“Music is a form of non-verbal language, and verbal language has musical components,” he said. “It involves the perception of different voice frequencies, intonations, beats and rhythms, for example.” 

Musical perception, he added, is directly linked to phonological awareness, a broad skill that includes identifying and manipulating units of oral language such as words, syllables, and phonemes.

“Phonological awareness is a basic skill that precedes reading,” Moreira said. “There appears to be a cascade effect: the better the musical perception, the easier it is to perceive different sounds; this skill in turn helps with reading single words, and from there, it’s easier to learn contextual reading and understand what’s written. This entire chain affects school performance.”

In Moreira’s view, the chief advantage of using musical perception as a predictor of reading and writing ability is that music is a universal language regardless of mother tongue and culture. The same methodology can therefore be used to test children all over the world.

M-factor

Dissatisfied with the tests previously available to measure musical perception, the group led by Moreira decided to create a new tool that could assess children’s ability to perceive variations in seven musical domains: timbre, meter, scale, loudness, contour, duration and pitch. 

To do this, they recorded a large number of short musical excerpts using sophisticated composition programs, with sounds of high quality and different musical timbres (including noise). The sounds encompassed a range of musical comprehension that extended well beyond the violin and piano excerpts on which such tests have traditionally focused.

All the sound stimuli were new compositions produced by Caio Giovaneti de Barros, a PhD student at the Arts Institute of São Paulo State University (IA-UNESP) and first author of the article.

Organized in pairs, these stimuli made up a total of 80 tasks that challenged children to decide whether each musical pair was the same or different in all seven domains.

Click on the links below to open the sound files (.wav format)

Two sound stimuli with similar pitches

Similar contours

Different durations

Similar scales

Different loudnesses

Different meters

Different timbres

“Existing tests used different combinations of musical domains. Some included only three, such as meter, pitch and rhythm. Others included four or five. Generally speaking, however, the sound quality of these musical stimuli was poor, consisting of MIDI files. Moreover, many tests used well-known tunes such as Happy Birthday to You. We made a point of creating entirely new musical stimuli so that the child’s auditory history wouldn’t influence the outcome,” Moreira said.

The test was applied to a random sample of 1,006 children aged 6-13 years from 14 schools in São Paulo State and equally distributed in the first to fifth grades. Students from private schools accounted for 38% of the sample.

The group analyzed the results statistically using a bifactor model (or general-specific model), assigning a different weight to each of the domains assessed.

“To our surprise, we found that measuring each domain in isolation told us very little about an individual’s musical perception,” Moreira said. “They all had very little weight compared to a general factor that appeared to govern the rest. We called this general factor the musical perception factor, or m-factor.”

The researchers compared the m-factor identified in the study with the g-factor put forward in the early twentieth century by English psychologist Charles Spearman as a general factor that governs an individual’s intelligence level. 

“In other words, it’s not enough for a child to have a good perception of meter alone,” Moreira said. “M-factor scores genuinely do reflect the quality of musical perception.”

The vast majority of the children tested in this study achieved average scores.

Chart showing the distribution of the m-factor in the population studied. Most achieved average scores. Left: musical perception scores well below average; right: musical perception scores well above average.

The difference between boys and girls was not statistically significant. This finding contrasts with those of previous studies pointing to a better performance by girls in verbal language tests.

Children enrolled at private schools performed slightly better than children from public schools. When individuals of different ages were compared, however, there were no significant differences. According to Moreira, this suggests that the m-factor does not improve with age.

“If the scores of twelve-year-olds aren’t above 6 on average, all the signs indicate that this musical perception is innate. It can be stimulated, but basically, it’s there from birth,” Moreira said. “However, to be sure about that, we’d have to do a follow-up study and assess the same children as they develop.”

The next step in the investigation, he added, is to apply the test to another group of children and try to correlate the m-factor scores with their school performance in reading and writing.

To be used as a tool for prediction in preschool children, however, the test would have to be adapted. “With 80 tasks, it takes too long for children under 6,” Moreira said. “If we can adapt it and show that it’s capable of predicting language acquisition difficulties, then it could be a basis for very early interventions such as music education classes, for example.”

The article “Assessing music perception in young children: evidence for and psychometric features of the m-factor” can be read at: journal.frontiersin.org/article/10.3389/fnins.2017.00018/full.