While dyslexia is most often classified as a reading disorder, it is also well known to affect how individuals process spoken language. Even in infancy, people at high risk for dyslexia seem to have difficulty processing speech sounds, according to Iris Berent, professor of psychology in the College of Science at Northeastern.

While more than 5 percent of the global population suffers from dyslexia, Berent said much remains unclear about what causes this disorder because we don’t completely understand how the brain decodes printed language. There are also mysteries surrounding how the brain deals with spoken language, but linguistics tells us it involves at least two different systems: the phonetic system and the phonological system.

“The phonetic system extracts the distinct building blocks from continuous acoustic sound,” Berent said. She further explained that the phonological system takes those blocks and builds patterns with them.

“Think of the metaphor of Lego blocks,” she continued. “The phonetic system gets the Legos from the plastic stuff; the phonological system builds patterns with them.”

Researchers have long believed that the phonological system was impaired in people with dyslexia. Yet, surprisingly, very few studies ever bothered to check. In a new paper in the journal PLOS ONE, Berent and her colleagues — Vered Vaknin Nusbaum of the University of Haifa, Evan Balaban of McGill University and Albert Galaburda of Harvard Medical School — show that the phonological system of dyslexics is intact. It is actually the phonetic system that is failing, she said.

To test the phonological system, Berent and her team presented both skilled and dyslexic readers with the sounds of fake words in Hebrew, the language used in the study. Of those fake words, some had sound patterns that are possible in Hebrew while others are not. The two groups were equally skilled at distinguishing between them, indicating that dyslexic participants could just as easily identify the Lego patterns as those without dyslexia.

Given the existing literature, this was extremely surprising. Indeed, when Berent designed this research, she said she “fully expected to validate the phonological hypothesis.” But the results showed no hint of a phonological deficit.

Instead, she found a host of subtle problems in the perception of speech sounds. For example, in one experiment, dyslexic individuals had a harder time distinguishing real words from fake words. A second experiment showed that dyslexics had a hard time distinguishing real human speech from digital sounds mimicking speech.

“So maybe it’s getting the Lego blocks, not patterning the Lego blocks, that is impaired,” she said. That is, the problem seems to reside in their phonetic systems.

To test this hypothesis, they also asked listeners to distinguish between discrete acoustic sounds (such as “ba” or “pa”). Once again, “there were some reliable differences in the two groups in identification, and even more so in discrimination,” she said.

In the past, dyslexia researchers accepted what Berent calls a “mushy definition” of phonology. “Some researchers identify phonology as any process related to speech processing, whether it is speech perception, or the mapping of letters to speech sounds,” she said. “I think the contribution of our work, is saying, ‘Look at the linguistics, look at what the two systems really are doing in human languages and maybe that will help you understand dyslexia.’”

Berent’s findings suggest that the disorder may reside in a lower-level component of speech perception, such as the auditory system. Other research suggests these difficulties might arise in the early development of the human brain, perhaps even before birth.