|
NARSAD Scientist Advances Explanation for Disordered Thinking in Schizophrenia
(Boston, MA -
) — According to NARSAD scientist Dean F. Salisbury, Ph.D., problems at two distinct stages of language processing may explain the disjointed thinking seen in schizophrenia. By tracing individuals’ brainwaves as they read words with multiple meanings, he and his colleagues found that people with schizophrenia made inappropriate associations with weakly related concepts, but memory deficits later made them default to the typical meaning of words even when the context pointed to an alternative meaning.
People with schizophrenia often veer off on tangents during conversations, confuse figurative with literal expressions, and lose their train of thought mid-sentence. Studying how they process language has occupied Dr. Salisbury, of Harvard Medical School and McLean Hospital, and his colleagues for years. “Thought disorder may be related to a singular cognitive abnormality, but it is not known where in the information processing stream the abnormality occurs or how it arises,” the 1995 and 1997 NARSAD Young Investigator wrote in the April issue of Clinical EEG and Neuroscience.
Ambiguity equals opportunity
To track that processing stream, scientists study responses to homographs, words that carry more than one meaning, like “palm”. Some experts think that deciphering meaning involves searching networks of nodes that dwell in memory, where each node represents a concept. Perceiving a word seems to spread activation among nodes representing related concepts along paths that are guided by contextual cues. In an article about California landscaping, people associate “palm” with a kind of tree and, more distantly, with plants; in the presence of a fortune-teller, they make different associations and comprehend a different meaning.
This process seems to break down in people with schizophrenia, causing misinterpretations that scientists can probe for hints to how the illness hobbles the brain. “Schizophrenic error patterns are likely not random,” Dr. Salisbury wrote. Although controversial, some theories suggest a “semantic bias” in people with schizophrenia that favors the dominant, or most common, interpretation of a word even when its clashes with contextual cues, causing them to misconstrue “toast” as browned bread even in a conversation about a wedding reception. This could stem from overactivation within the network. Other theories blame faulty higher-order processes such as attention.
In a series of studies by Dr. Salisbury and colleagues, participants with and without schizophrenia read sentences, each of which contained a homograph as well as a related or unrelated word. Meanwhile, electrodes recorded a brainwave called the N400 event-related potential to see if priming had occurred. In priming, a conceptually related word that precedes a homograph speeds its processing; for instance, “lime” primes “lemon.” As Dr. Salisbury explained, “The N400 is smaller when priming has occurred.”
In one study by Dr. Salisbury’s research team, participants read four-word sentences. Each included a homograph and ended with a word that clarified its meaning, with a long interval between them. Compared to the control group, participants with schizophrenia showed greater N400s to sentences that conveyed less typical meanings. According to Dr. Salisbury, a semantic bias kept them from correctly gauging the meaning from the context.
People without schizophrenia tend to respond in a graded way to these kinds of tasks, with N400s that range from large for unrelated words, to midsize for rare associated meanings, to small for strong associates. In contrast, the study found that people with schizophrenia responded to all sentence endings with a larger N400. That could reflect a failure to pay attention, but a follow-up study ruled out that explanation, leading the researchers to suspect poor processing of contextual information.
Activation versus memory
Another study asked participants to decide whether two words were related. To detect where language processing breaks down, it varied the time between the homograph and cue word. Shorter intervals reveal the early spread of activation among concepts; longer ones offer a glimpse of verbal working memory.
At quick presentation rates, people with schizophrenia showed a reduced N400 to unrelated words, indicating inappropriate priming of homographs by distantly related contexts. At longer intervals, they showed larger N400s than controls regardless of how closely the words were related. Dr. Salisbury attributed this to problems with holding contextual information in verbal working memory.
Taken together, these findings point to double trouble in schizophrenia. Early in language processing, excessive priming spreads activation farther than normal through the network. After that, item representations in working memory fade so fast that, after one second, Dr. Salisbury wrote, “only the most strongly associated concepts will remain in the buffer,” creating a bias that hampers the use of contextual cues. These conclusions set the stage for further efforts to clarify the memory deficits that underlie thought disorder in people with schizophrenia.
 
|
Media Contact
Upcoming NARSAD Events
Latest News from NARSAD
Spotlight
|