Studies have shown that computer analysis of a person’s speech can reveal motor control problems, indicating deficits in comprehending language, cognition and decision-making abilities. Dr. Philip Lieberman is using these findings to develop a system that monitors speech patterns to predict mental abilities. The research is aimed at a system that will monitor the flow of conversation, and Lieberman is studying the speech and behavior patterns of Mount Everest climbers to develop and verify the program that could be used to monitor spacecraft crews. The linked speech production, cognitive and language deficits appear to reflect impairment to brain structures implicated in Parkinson’s disease. Lieberman and his colleagues have applied similar techniques to study Parkinson’s disease patients and to evaluate new treatments.
Overview
Speech Monitoring Cognitive and Personality Alterations
Principal Investigator:
Philip Lieberman, Ph.D.
Organization:
Brown University
Technical Summary
Deep-space missions will expose crews to cosmic rays, which may damage brain structures including the basal ganglia and hippocampus. These subcortical structures are linked with cortical regions in circuits regulating motor control, cognition and personality. As models for crews suffering neurological damage, we use climbers ascending Mount Everest and patients with Parkinson's disease (PD). At altitudes above Everest Base Camp, the low oxygen content can damage the principal basal ganglia output structure, the globus pallidus. The hippocampus also is sensitive to hypoxia. The ascent of Everest also resembles long-term space flight in that a small group in close contact for an extended period must make critical decisions in stressful, life-threatening situations.
PD, which affects basal ganglia function, impairs motor control but also causes deficits in cognition and language comprehension more severe than those in Everest climbers. Climbers and PD patients offer complementary Earthbound space-analogs for milder and more extreme forms of the dysfunction astronauts may suffer on long-term missions. The rationale for using speech measures as cognitive indices is that the basal ganglia regulate sequences of speech motor acts as well as of cognitive operations. We can expect the corresponding circuits to show parallel declines in function due to hypoxia, PD or radiological damage. One index we focus on is voice-onset time (VOT). Stop consonants in word-initial position are produced by first closing the lips or placing the tongue against the roof of the mouth. The obstruction then is abruptly released, yielding a burst with distinct acoustic properties. Speakers must also adjust the larynx to produce a vowel after the burst. For voiced stops, like b, the time between burst and vowel, the VOT is less than ~30 msec, while for voiceless stops, like p, it is greater than ~50 msec. Hence, VOT production requires precise control of the sequence of oral and laryngeal actions.
PD patients show reduced differences between VOTs for voiced and voiceless stops. Such VOT convergence correlates with deficits in sentence comprehension. On Everest, we obtain measurements of VOT and other speech parameters, e.g., vowel duration, by using radios to record climbers reading words at different altitudes. We have measured cognition by various tests. Climbers carry picture booklets so that we can administer, by radio, sentence-picture matching tests of language comprehension. To measure cognitive flexibility, we give the Odd-Man-Out test (OMO) and, in 2004, the Wisconsin Card Sorting Task. These require subjects to select stimuli using one of several possible criteria and then shift criteria following feedback. In 2003, we added the Mini-Cog test battery developed by Stephen Kosslyn's NSBRI group at Harvard. This battery, which tests cognitive capacities including verbal and spatial working memory, is implemented on Palm Pilot PDAs that climbers carry.
In 2004, a group of climbers also took an implicit contextual learning test that reflects hippocampal function. Results over successive expeditions bear out our hypotheses. At high altitudes, many climbers show VOT convergence and increased vowel duration. Crucially, these changes track impairment on cognitive tasks that depend on basal ganglia function. Greater VOT separation decreases were seen for larger increases in sentence comprehension response time (RT) with altitude. In 73% of cases, the presence (or absence) of reduced VOT separation signaled the presence (or absence) of deficits in comprehension RT or OMO performance. VOT separation had a similar 71% hit rate in discriminating impaired from unimpaired performance on Mini-Cog tests of working memory or vigilance. Vowel duration had a hit rate of 85% for sentence comprehension or OMO and was significantly correlated with verbal working memory RT.
Our speech metrics track severe cognitive deficits as well. One climber developed profound speech motor sequencing deficits and similarly extreme deficits in cognitive flexibility. Advised of his condition, he nonetheless proceeded to climb upwards, alone, through a storm; two days later he fell to his death. Preliminary analyses indicate that hypoxia impaired hippocampal function. Like amnesics, climbers at higher altitudes failed to show implicit contextual learning. Altitude did not affect performance on tasks that do not place demands on basal ganglia or hippocampal circuitry. Our PD studies support our hypotheses on the basal ganglia's role in understanding language.
By tracking eye movements as subjects do a sentence-picture matching task, we have shown that patients with high OMO error rates also exhibit impaired processing at points in sentences that require linguistic set-shifting. Our techniques may be used to monitor PD and have potential for detecting early memory loss in Alzheimer's disease, enabling interventions that could slow further decline. We are exploring the assessment of verbal apraxia in children, a disorder affecting linguistic and cognitive development that may result from hypoxia during birth. Our techniques have potential applications in general aviation, where hypoxia has led to disasters. In this project's next phase, we will apply speech analysis techniques to study task-induced stress in synergy with Kosslyn's group and David Dinges' NSBRI group at the University of Pennsylvania School of Medicine.
Earth Applications
Such early detection would permit clinicians to take maximal advantage of therapies, now under development, that can delay or even arrest further decline. In addition to Parkinsons disease, our techniques may have application to the diagnosis, assessment and treatment of other human pathologies stemming from impaired dopaminergic basal ganglia function in neural circuits regulating speech production, cognition and personality. These include not only neurodegenerative diseases but also the results of acute insult. For example, hypoxia during birth can lead to verbal apraxia in children a syndrome where speech motor and orofacial motor control is degraded and which can result in cognitive and linguistic deficits.