The processing of negative sentences: neural and behavioral correlates in healthy volunteers and brain-damaged population

All and only human languages are equipped with the possibility of negative predications, but the neural and cognitive correlates of this ability are still unknown. In particular, it is not clear whether negative sentence processing is intrinsically more complex than affirmative sentence processing or not. On one hand, longer RTs and lower accuracy observed with negative compared to affirmative sentences in behavioral tasks have been traditionally linked to greater computational demands, although recent studies reported cases where the behavioral patterns of affirmative and negative sentences are aligned. On the other hand, the neurofunctional results collected so far do not allow establishing if negative sentences actually recruit greater computational resources than affirmative sentences. In particular, previous functional Magnetic Resonance (fMRI) studies reported that, during sentence-verification tasks, negative sentences compared to affirmatives are associated with increased BOLD signal in left hemisphere language areas. However, when the effort imposed by the task is limited as in passive listening or lexical decision paradigms, this pattern is not observed. The current work integrates the research on linguistic negation with two studies that explore the neural and behavioral correlates of affirmative and negative sentence verification in healthy volunteers and brain-damaged population. The first study tested nineteen healthy subjects with a sentence-picture verification paradigm in a fMRI experiment. Increased activation in the left inferior frontal gyrus was found with negative vs affirmative sentences, but also with false vs true affirmative sentences. These results suggest that negative sentences are not intrinsically more complex than affirmatives, but are at a disadvantage during tasks that involve competing semantic representations. The second study tested t hirty-one individuals with aphasia and thirty-seven individuals with diagnosed degenerative dementia with an adaptation of the sentence-picture verification paradigm adopted in the first study. The comparison between patients and the control group (n=50) revealed that sentence- picture verification is overall problematic for people with brain injury, especially if sentences contain negation. Moreover, in focal patients the sentence-picture verification test allowed identifying distinct cognitive domains that may be selectively interested by the damage. We propose that the complexity engaged by linguistic tasks with negative sentences has not to do with negation processing per se, but with domain-general cognitive mechanisms that go beyond pure language processing.