Lemons and Trust: the Contribution of Anesthesia to the Study of the Neural Substrates of Concrete and Abstract Word Processing

There is evidence that abstract and concrete words are represented and processed differently in the brain. Numerous studies suggest the activation of a bilateral network for both abstract and concrete words, with a greater involvement of posterior, sensory areas in concrete word processing, and a more focal activation of anterior regions, involved in verbal processing, for abstract words. This Ph.D. thesis aimed at investigating the different neural substrates of concrete and abstract words by studying memory priming during general anesthesia. Implicit memory tasks, in fact, seem to be immune to the concreteness effect and recent neurophysiological studies suggest that conscious and unconscious semantic activation involve similar brain areas. Experiment 1 focused on the priming effect for intraoperatively primed abstract and concrete words in patients under general intravenous (propofol) anesthesia. Considering the specific brain targets of propofol, I hypothesized a stronger priming effect for concrete than for abstract words. Implicit memory for primed words was tested with a three-letter word stem completion test, in which half of the stems referred to primes, and half were foils. Both stimulation and testing were auditory, to avoid cross-modality interference. A control group of patients, who did not receive any intraoperative stimulation, but completed both concrete and abstract word stem completion test, was also recruited. As expected, a priming effect was found for concrete words, since the number of target hits was significantly higher than the number of non target hits. This difference did not apply to abstract words. The abstract experimental group performed comparably to controls. These results support the thesis that abstract word processing relies on the activity of anterior brain areas, as for example the inferior frontal cortex, which are suppressed by propofol. The results would also confirm that priming, investigated through word stem completion, is not a simple perceptual, pre-semantic task, but engages multiple processes, including semantic access. As suggested by electrophysiological studies, semantic access might occur at very early stages of verbal processing, thus explaining a selective intraoperative priming effect for concrete words only. To further investigate these hypotheses, a second experiment was conducted. In experiment 2 the same methodology of experiment 1 was applied, but the volatile anesthetic sevoflurane was used, which is known to lower activity in brain regions located more posteriorly than those suppressed by propofol. The most interesting finding was the presence of a priming effect also for abstract words, which would confirm data from neuroimaging studies of a greater engagement of anterior brain regions in abstract word processing. To better define the involvement of the frontal cortex in the processing of abstract words, a series of patients undergoing awake surgery for brain tumor removal was studied in experiment 3. Patients performed a lexical and a semantic decision task, together with a standard intraoperative cognitive monitoring, during direct cortical stimulation. The fundamental role of the left inferior frontal gyrus in abstract word processing was confirmed.