It is obvious from observing deaf children that dysfunction in the sensory registration of the auditory signal is sufficient to disrupt speech perception and give rise to a global language disorder. That deafness is also likely to limit the acquisition of reading, even in individuals who are fluent in a visual sign language, attests to the importance of auditory/phonologic aspects of learning to read. Although children with LLI by definition have normal peripheral hearing, processing of the complex acoustic signal involved in speech perception goes far beyond the level of peripheral sensation. In addition to their linguistic deficits, children and adults with LLI are characterized by central processing deficits, particularly those involved in processing brief, rapidly sequential stimuli differing in frequency or pitch (often referred to as temporal processing deficits). A question that has stimulated considerable research and debate in this field for more than 30 years is precisely what role these nonlinguistic processing deficits play in the origin of LLI.
In an attempt to investigate whether one or more central processing mechanisms may be impaired in some children with developmental language disorders, research has focused on investigating those higher level auditory perceptual processes that are likely to be most directly involved in speech perception. A comprehensive series of psychophysical studies using a variety of nonlinguistic acoustic signals have focused attention on the temporal and spectral processing abilities of individuals with LLI. One of the most replicated findings in this field is that individuals with LLI, as a group, are significantly impaired in their ability to track rapid (tenths of a millisecond) spectral changes that occur within as well as across acoustic signals.
It has been hypothesized that the individual phonemes (speech sounds) that are the building blocks of oral languages must be learned from environmental exposure, using basic Hebbian learning principles. According to Hebb, "ensembles of neurons that fire together wire together", and repeated exposures over time sharpen the precision of cortical sensory maps. It has also been hypothesized that individual differences in processing speed will affect the time needed to integrate sensory input, in turn affecting the sharpness or precision of categorical representation of phonemes in the nervous system. If this were the case, then individuals with slow auditory processing would be expected to have less fine-grained representations of the phonemes of their language, leading to considerable difficulty discriminating the fine-grained, rapidly changing acoustic differences (such as formant transitions) that characterize many phonemes.
This hypothesis has been supported by studies using a variety of behavioral, neuroimaging, and remediation techniques. Take as an example the acoustic spectrogram representing the frequency changes that occur over time as two consonant-vowel (CV) syllables, such as the /ba/ as in body and the /da/ as in dot are produced. These speech syllables are characterized by rapidly changing acoustic cues. Importantly, / ba and /da/ share the same acoustic signature during the majority of the signal, the time over which the common vowel is being produced. As such, these syllables must be discriminated based primarily on the initial, brief (approximately 40-msec), transitional component of the signal that is created as the speech articulators move from their initial place of articulation into the steady-state portion of the vowel. This is true of many consonants whose acoustics change exceedingly rapidly as we produce them within word contexts.
It was hypothesized that LLI children, because of their slow auditory processing, would be significantly impaired in their ability to discriminate between syllables characterized by rapid acoustic changes. This hypothesis was supported by the results of experimental tests. Further studies showed that the limiting factor underlying the inferior speech processing performance of individuals with LLI is indeed the duration of brief or rapidly successive acoustic changes within syllables and words. Significant improvement in speech discrimination was found when the time over which these acoustic cue differences occurred intrasyllabically was increased. In these experiments, computer-synthesized speech was used to extend the duration of the initial transitional acoustic component of CV syllables from 40 to 80 msec. The ability of children with LLI to discriminate between syllables that incorporated longer duration transitions was improved significantly over their ability to discriminate between syllables with usual-length transitions.
This finding has led to the development of new intervention programs called Fast ForWord ( www.fastforword.com ) that are now used widely in schools throughout the US as well as internationally for treating individuals with LLI as well as children learning English as a second language. A computer algorithm was developed that differentially extends and enhances the amplitudes of most rapid acoustic changes occurring within natural speech. This acoustically modified speech is then used in a series of computer training exercises designed to train the individual in various phonological processing, language comprehension, and reading decoding tasks. The computer exercises incorporate neuroplasticity-based learning routines (based on Hebbian learning principles) that individually adapt to each participant's trial-by-trial performance. As the individual progresses in the language and reading exercises, the amount of acoustic modification decreases (with the goal being normal speech). Another exercise specifically focuses on increasing the speed of auditory sequential processing.
These novel computerized training approaches, based on the results of years of cognitive neuroscience research, have proven highly effective in improving the clinical and educational outcomes for individuals with a variety of oral and written language learning problems, including the reading abilities of dyslexics. Similar training programs within the Fast ForWord series have been developed to improve selective attention, memory and other foundational cognitive skills essential to efficient learning that are proving to be helpful to children with a variety of cognitive deficits, including Attention Deficit Disorders, Central Auditory Processing Disorder and Autism. Recent neuroimaging studies using functional magnetic resonance imaging (fMRI) have demonstrated significant correlations between amelioration in deviant metabolic activation in language structures of the brain and improvement in language and reading abilities after the use of these training programs.