Understanding and Treating Speech Sound Processing Problems in Autism

Project Summary

Although individuals with autism are known to have significant communication problems, the neural mechanisms responsible for impaired communication are poorly understood.We are working with an animal model of autism to identify a potential cause of speech sound discrimination impairments and quantify the beneficial effects of two common autism therapies: auditory training and environmental enrichment. A better understanding of these mechanisms may aid the design of improved behavioral and sensory therapies to reduce communication impairments in autism.


Children with autism have difficulty communicating with others.

It is estimated that 1 in 150 people in the United States is affected by autism (Newschaffer et al., 2007). Autism is characterized by social impairments, language deficits, and repetitive behaviors (Rapin, 1997). Detailed psychophysical studies have shown that autistic individuals are severely impaired in their ability to process the subtle cues used in everyday communication and social interactions (Rapin and Dunn, 2003; Siegal and Blades, 2003; Alcantara et al., 2004).

Abnormal perception of speech sounds appears to exacerbate the communication and social impairments characteristic of autism. Recent functional imaging studies have confirmed serious deficits in speech sound processing in both children and adults with autism (Boddaert et al., 2004; Bomba and Pang, 2004). Autistic children exhibit impaired processing of speech sounds, but not tones (Ceponiene et al., 2003; Kuhl et al., 2005; Oram Cardy et al., 2005; Whitehouse and Bishop, 2008). These impairments likely arise from a degraded neural representation of speech sounds. Less activation is seen in the left hemisphere in autism in response to speech sounds (Boddaert et al., 2003), and this hemisphere is thought to be important in processing the rapid temporal information in speech (Hickok and Poeppel, 2000). Speech sound processing impairments do not improve over time, and are still seen in adults with autism (Buchwald et al., 1992; Bomba and Pang, 2004; Gervais et al., 2004). Unfortunately, the poor spatial and temporal resolution of current imaging methods obscures the neural basis of the impairment. A better understanding of the neural representation of speech sounds in autism would be useful in developing more effective interventions to improve speech and language development.

Prenatal valproic acid exposure leads to developmental delays and autism.

Although genetic and epidemiological studies have demonstrated that autism does not have a single cause, inappropriate regulation of neural plasticity appears to be a major contributor to each of the diagnostic behavioral deficits. Mutations in a wide range of genes predispose children to autism, especially genes associated with synaptic plasticity and arousal mechanisms. Genetic studies have revealed that mutations in genes associated with neural specification, growth cone guidance, synapse formation, neuromodulator release or cell signaling can predispose children to autism (Cook, 2001).

Valproic acid (VPA) is an anticonvulsant and mood-stabilizer used to treat epilepsy, migraine, and bipolar disease under the trade names Depakote and Stavzor.Prenatal exposure to VPA greatly increases the likelihood of developmental delays (Ardinger et al., 1988; Moore et al., 2000; Viinikainen et al., 2006).Many of these children require speech therapy or educational support (Moore et al., 2000; Viinikainen et al., 2006).†† Prenatal exposure to VPA is also strongly associated with autism (Christianson et al., 1994; Williams et al., 2001; Ornoy, 2009). It has been estimated that approximately 10% of VPA exposed children develop autism, compared to the less than 1% risk in the general population (Moore et al., 2000; Rasalam et al., 2005).

Rats exposed to a single high dose of VPA in utero exhibit many of the classic neural abnormalities and behavioral deficits observed in autism, including brainstem defects, loss of cerebellar neurons, serotonergic abnormalities, impaired social interactions, increased repetitive behaviors, enhanced anxiety, locomotor hyper-activity, abnormal fear conditioning, lower sensitivity to pain, higher sensitivity to non-painful sensory stimulation, impaired pre-pulse inhibition, and enhanced eye-blink conditioning (Schneider et al., 2001; Schneider and Przewlocki, 2005; Markram et al., 2008). Recent brain slice studies indicate that somatosensory cortex, prefrontal cortex, and amygdala neurons in VPA exposed rats are hyperexcitable and hyperplastic (Markram et al., 2008; Rinaldi et al., 2008b; Rinaldi et al., 2008a; Silva et al., 2009).These results indicate that in utero VPA exposure in rats is an appropriate animal model of autism (Markram et al., 2007a).

Autism therapies likely work by stimulating neural plasticity, but the evidence is weak.

Speech training and enrichment are two widely used therapies to reduce the symptoms of autism (Lovaas, 1987; Kasari et al., 2008; Pickett et al., 2009).Both treatments have been hypothesized to generate beneficial forms of neural plasticity (Burack et al., 2001; Dawson, 2008). However there is no direct evidence of therapeutic plasticity in autism.

Recent evidence suggests that intensive behavioral training can substantially improve language proficiency in children with language deficits (Merzenich et al., 1996; Tallal et al., 1996; Kujala et al., 2001; Habib et al., 2002).More than 300,000 learning-impaired children have been trained on computer-based games and exposed to modified speech in an attempt to improve their language ability (Merzenich et al., 1996; Tallal et al., 1996; Tremblay et al., 1998; Kujala et al., 2001; Habib et al., 2002; McCandliss et al., 2002).While considerable controversy remains, some studies have reported improvements as large as two grade levels after three weeks of training (Merzenich et al., 1996; Tallal et al., 1996). Anecdotal results indicate computer based speech therapy is also effective in autistic children (Tallal, 2000). Although some of these treatments are based on neuroscience research, many of the most basic questions about the neurological basis of speech comprehension impairments and rehabilitation remain unanswered (Kraus et al., 1998).

Both humans and animals substantially improve their performance on almost any task after several weeks of practice.In some cases, the improvement is specific to the physical features of the training stimuli, and in other cases improvements generalize to novel stimuli (Fahle, 2005).Physiological studies in awake and anesthetized animals have shown that training with simple stimuli can generate receptive field plasticity that is specific to the region of the sensory map engaged during training (Buonomano and Merzenich, 1998) (but see (Brown et al., 2004; Ghose, 2004)).More recent behavioral studies in rats and primates have shown that temporal response properties can also be altered by operant training (Beitel et al., 2003; Bao et al., 2004).These results support the general hypothesis that intensive training can generate cortical plasticity, but these results are insufficient to establish what forms of plasticity would most benefit speech processing.

Studies of plasticity in speech impaired subjects typically train on multiple contrasts in hopes of maximizing the generalization of the training effect.Eight weeks of training on phonological awareness, auditory processing, and language processing skills through interactive games (Earobics, Cognitive Concepts, Inc., Evanston, IL) enhanced cortical evoked responses inchildren with learning problems (Diehl, 1999; Hayes et al., 2003).The reliability of cortical response timing also increased with training (Warrier et al., 2004).Auditory brainstem responses were not altered.Several months of experience with a cochlear implant reduces P1 latency in deaf children by 50 to 200 msec (Sharma et al., 2005).Collectively, these results demonstrate the promise of targeted behavioral therapy to facilitate functional recovery. Understanding the spatiotemporal pattern of neuronal activity that is evoked by such training may facilitate the design of new remediation strategies for communication deficits in individuals with autism (Pickett et al., 2009).

Environmental enrichment significantly improves outcomes in a variety of neurological deficits and is commonly recommended for autistic children (Will et al., 1977; Kolb and Gibb, 1991; Hannigan et al., 1993; Rampon et al., 2000; Hockly et al., 2002; Morley-Fletcher et al., 2003; Jankowsky et al., 2005).Molecular, cellular, and systems studies have documented positive effects of environmental enrichment.Enrichment increases synaptic strength and density in the cerebral cortex (Volkmar and Greenough, 1972; Globus et al., 1973; Greenough et al., 1973; Sirevaag and Greenough, 1987; Nichols et al., 2007). Enrichment also causes cortical neurons to respond more strongly, more selectively, and more quickly to sensory stimuli (Beaulieu and Cynader, 1990; Coq and Xerri, 1998; Engineer et al., 2004; Nichols et al., 2007).Environmental enrichment is sufficient to reverse many of the behavioral abnormalities in rats prenatally exposed to VPA after (Schneider et al., 2006).

In research supported by the Cure Autism Now Foundation, we documented that enrichment improves temporal response properties in primary auditory cortex (A1).Specifically, we observed faster onset latencies and increased paired pulse depression that could serve to improve speech processing in autism (Engineer et al., 2004; Percaccio et al., 2005).However, despite the widespread use of enrichment and speech therapy to treat autism, no human or animal study has documented the effect of these interventions on neuronal activity induced by speech sounds.

Justification of the Model Species

Though animals are clearly incapable of understanding language, there is now considerable evidence that many species, including rats (Clark et al., 2000; Reed et al., 2003; Toro et al., 2003; Toro and Trobalon, 2005; Toro et al., 2005; Eriksson and Villa, 2006; Floody and Kilgard, 2007), are able to reliably discriminate human speech sounds (Kuhl and Miller, 1975, 1978; Kuhl, 1986; Sinnott and Adams, 1987; Dooling et al., 1989; Kluender and Lotto, 1994; Fitch et al., 1997; Pepperberg, 1999; Sinnott and Mosteller, 2001; Ehret and Riecke, 2002; Kaminski et al., 2004).The observation that animals exhibit categorical perception of speech sounds (Kuhl, 1986) suggests many of the basic neural mechanisms used to distinguish speech sounds might be conserved in mammals (Kuhl, 1986; Kluender, 2002).Recent reports suggest the early stages of speech sound processing in humans and animals are similar (Steinschneider et al., 2005; Engineer et al., 2008).The key advantages of an animal model are 1) the potential to precisely control the environmental and genetic factors, and 2) the potential to directly record individual neurons with millisecond precision.

Our studies will be conducted in rats because this important model system offers numerous experimental advantages over other animal models.The brain anatomy, physiology, and behavior of rats are the best studied of any animal species.More auditory training studies have been conducted in the rat than in the cat, monkey, and ferret combined. The ratís small size and reasonable cost makes it possible to use a larger number of subjects than would be possible with many other species.The wealth of experimental techniques available and the long history of neuroscience research on rats suggest this species is a logical choice in which to investigate the basic neural mechanisms that could contribute to speech processing impairments in humans.

Scientific Summary

††††††††††† The poor resolution of human imaging techniques obscures the neural basis of speech processing impairments, so we propose to evaluate speech sound coding in the valproic acid (VPA) animal model of autism, and quantify the beneficial effects of two common autism therapies: auditory training and environmental enrichment. Speech sounds evoke specific spatiotemporal patterns of cell firing in the central auditory system of normal rats (Engineer et al. 2008). The first aim of the project is to determine the consequence of VPA exposure on the collicular and cortical representations of speech sounds. Our preliminary results indicate that in utero VPA exposure severely degrades the precise spatiotemporal patterns evoked by speech sounds in auditory cortex. As in autism, the longer latency in our animal model is significantly greater for speech sounds compared to tones. The second aim of the project is to determine the behavioral consequences of VPA exposure on speech sound discrimination. If the neural spatiotemporal representations of speech sounds are degraded, then it is possible that certain speech sounds may not be distinguishable in VPA treated rats. We therefore predict that speech sound discrimination will be impaired in VPA exposed rats. The third aim is to determine the effects of speech training and environmental enrichment on speech evoked activity in VPA exposed rats. Based on previous studies, we predict that both speech training and environmental enrichment will relieve the degradation of the cortical responses to speech sounds and restore speech sound discrimination to control levels in VPA treated rats. The results of the proposed studies will add to our understanding of the neural mechanisms that are associated with speech sound coding. Insights derived from these studies may influence the development of new behavioral and sensory techniques to treat the communication impairments in autism that result in part from degraded speech sound discrimination.

Research update

Return to Kilgard Lab website