Prefrontal activation during verbal fluency tests in schizophrenia—a near-infrared spectroscopy (NIRS) study

Abstract

Letter- or semantically-cued verbal fluency tests (VFT) induce different, but overlapping activities in the prefrontal cortex (PFC) in healthy subjects. Functional differences between letter and semantic VFT may be important in elucidating the nature of language-related problems in schizophrenia since more profound impairment in semantically-cued retrieval may be suggestive of semantic system dysfunction specific to this disorder. However, the functional differences between the letter vs. semantic VFT on prefrontal metabolism in the subjects with schizophrenia have not been described. In the present study, we used near-infrared spectroscopy (NIRS) to assess hemoglobin concentration changes in the bilateral PFCs in subjects and controls. Sixteen subjects with schizophrenia and 19 healthy controls performed letter and semantic VFTs. Prefrontal hemodynamic changes were simultaneously monitored by NIRS. While healthy subjects performed both tasks equivalently, the subjects with schizophrenia showed more compromised performance in the semantic VFTs compared to the letter VFTs. NIRS measurement revealed that the pattern of PFC activation was greater during the letter VFT when compared to the semantic VFT in the healthy subjects, suggesting more prominent PFC involvement in letter-cued retrieval. In contrast, the subjects with schizophrenia showed the opposite pattern of activation, implying that the semantic mode of lexical access might impose greater cognitive demands on the PFC. The present study is the first to detect abnormal patterns of PFC activation in adults with schizophrenia in response to the distinct cognitive demands associated with letter and semantic VFT.

Introduction

Previous functional imaging studies have repeatedly observed activation in the prefrontal cortex (PFC) during a verbal fluency test (VFT), in which subjects were asked to generate words by either a cue of a given letter (letter VFT) or when given a semantic category (semantic VFT). Depending on whether retrieval was made based on a letter cue (e.g., F or A) or a categorical cue (e.g., vegetable or tools), overlapping but different patterns of activation have been observed. Letter VFT usually produced greater PFC activity than did semantic VFT. Retrieval of words on the basis of their initial letter is not a significant component of natural language production; therefore, unlike semantic retrieval, subjects need to employ an unfamiliar access route to the lexicon (Wood et al., 2001). In contrast, semantic VFT produces prefrontal activity as well as activity weighted towards the left temporal lobe (Gourovitch et al., 2000, Mummery et al., 1996), suggesting that retrieving words in response to category labels, rather than in response to letters, may depend more on semantic processing.

One of the core features of the psychopathology of schizophrenia is language-related problems. This is observed in both the linguistic and semantic domain. VFT performance in subjects with schizophrenia is of interest, because letter/semantic VFTs can examine phonological/semantic modes of lexical access independently. Several studies reported differential impairment between letter and semantic VFT in schizophrenia. In such studies, the impairment in semantic processing has been reported as being more profound (Goldberg et al., 1998, Gourovitch et al., 1996, Phillips et al., 2004), suggestive of dysfunction that is specific in the semantic mode of lexical retrieval. Recent meta-analysis of studies investigating the differential VFT performance in schizophrenia (Bokat and Goldberg, 2003) also supported this view. Recently, Curtis et al. (1999) reported that hypofrontality in subjects with schizophrenia might be specific in the letter-cued retrieval process: subjects with schizophrenia showed attenuated frontal activation during letter VFT; however, they did not show differences in activation compared to normal subjects while performing a semantic decision task. As mentioned above, in the case of healthy subjects, letter VFT produces greater increases in frontal activation. Therefore, it can be hypothesized that more profound abnormalities of semantic processing in schizophrenia might somehow be compensated by excessive recruitment of prefrontal neural circuitries.

In the present study, we used near infrared spectroscopy (NIRS) for assessment of PFC activation during VFT in subjects with schizophrenia. NIRS is an optical method that can be used to investigate the oxygenation of brain tissue (Jobsis, 1977). Near-infrared spectrum light can penetrate the skull and is absorbed mainly by two chromophores: oxygenated hemoglobin (oxy Hb) and reduced hemoglobin (deoxy Hb). Changes in oxy Hb and deoxy Hb concentrations ([oxy Hb] and [deoxy Hb]) can be calculated from the amount of absorbed near-infrared light using the modified Lambert-Beer Law. NIRS has been successfully used to detect prefrontal activation related to VFT in healthy populations (Herrmann et al., 2003, Toichi et al., 2004) or subjects with schizophrenia (Watanabe and Kato, 2004). Using NIRS, the most commonly reported finding is activation of PFC with an increase in oxy Hb and a small decrease in deoxy Hb, resulting in an increase in total hemoglobin concentration ([total Hb], i.e., [oxy Hb]+[deoxy Hb]). The increase in oxy Hb, coupled with the corresponding decrease of deoxy Hb can be considered a robust marker of cerebral oxygenation changes caused by cognitive activation.

On-line monitoring of prefrontal activity during overt word generation has been hampered by technical difficulties in conventional imaging techniques: fMRI requires strict motion restriction in order to minimize movement-related artifacts. In contrast, NIRS does not need such severe motion restriction. Small photo-sensors attached to the forehead measure the infra-red light that penetrates tissues, and the small body movements that are associated with speech have no significant effects on the measurement. Although spatial resolution is low, time resolution of NIRS is superior to fMRI. Therefore, NIRS makes it possible to monitor frontal activity during the time in which subjects generate spoken words. The NIRS technique is non-invasive and generally associated with only modest amounts of psychological stress associated with the measurement procedure. As NIRS can be performed in an office or even at bedside, it is well suited for studies in subjects with psychiatric disorders. In this study, we conducted repeated NIRS scanning throughout the task performance (e.g., every 2 s for 3 min in the present study). This allowed for real-time monitoring of prefrontal blood flow changes in response to the different modes of lexical access. Using a one-channel NIRS system, Watanabe and Kato (2004) previously reported attenuated activation of the left PFC during VFT in subjects with schizophrenia. To our knowledge, no study has directly compared the functional differences between letter and semantic VFT on prefrontal hemodynamics within the same subjects with schizophrenia. The present study is the first NIRS study that investigated bilateral prefrontal hemodynamics using this differential cognitive task design.