تعامل "متناقض" قشر شنوایی اولیه در بیماران مبتلا به توهم شنوایی کلامی: فرابررسی مطالعات کارکردی تصویربرداری

The existing literature on neuroimaging studies of auditory verbal hallucinations (AVHs) in patients with schizophrenia contains an apparent “paradox” in that the same areas in the auditory cortex seem to be both activated and deactivated in relation to AVHs, depending on whether an external auditory stimulus is present or not. We performed meta-analyses of neuroimaging studies examining patients with schizophrenia during the processing of auditory stimuli and in individuals experiencing hallucinations in the absence of auditory stimuli to examine whether the auditory cortex shows the paradoxical decrease/increase pattern across studies. Databases PubMed and ISI Web of Knowledge were queried with the combination of the keywords “auditory verbal hallucinations”, “auditory hallucinations”, “fMRI”, “PET”, “imaging”, yielding 11 studies involving comparison between schizophrenia and control group during external auditory stimulation, and 12 studies of hallucinating subjects experiencing AVHs and resting in the absence of auditory stimulation. The data were analyzed using Activation Likelihood Estimation method. The results showed overlapping increased activation in the absence of an external stimulus, and decreased activation in the presence of an external auditory stimulus in the left primary auditory cortex and in the right rostral prefrontal cortex, confirming the “paradoxical” brain activation in relation to AVHs. It is suggested that the “paradox” may be caused by an attentional bias towards internally generated information and failure of down- and up-regulation of the default mode and auditory processing networks, respectively, with the consequence that the spontaneous activation in the absence of an external stimulus shuts down the perceptual apparatus for further processing.

Auditory verbal hallucinations (AVHs) are subjective experiences of “hearing voices” in the absence of corresponding external auditory stimulation. A characteristic feature of AVHs is the perceptual quality of the “voice”, which is typically experienced as a real person speaking with distinct perceptual qualities like accent, emotional valence, and timbre, as if in direct interpersonal communication. AVHs are reported by patients with various psychiatric and neurological disorders, most commonly schizophrenia (Choong, Hunter, & Woodruff, 2007), but can also be present in non-psychiatric populations without any accompanying psychotic symptoms (Johns and van Os, 2001 and Sommer et al., 2008b). The fact that AVHs may occur in the general population suggests that AVHs should be studied not only in the context of a diagnostic category, but as an independent phenomenon which can present itself in a variety of situations (Hugdahl, 2009), in turn suggesting a dimensional view of schizophrenia (see e.g., David, 2010). The similarities between experiencing AVHs and perceiving real auditory stimuli highlight the necessity of understanding the role of the auditory cortex in AVHs. Neuroimaging studies of hallucinating patients while they are experiencing AVHs suggest that auditory cortex is showing increased activation compared to alternating periods when no hallucinations are present (Dierks et al., 1999, Shergill et al., 2000 and van de Ven et al., 2005). This finding could be interpreted as increased spontaneous neuronal activity, or hyper-excitability of the auditory cortex. Spontaneous fluctuations of activation in sensory cortices have been shown to affect stimulus processing (Boly et al., 2007, Hesselmann et al., 2008 and Northoff et al., 2010), with increased pre-stimulus activation in auditory cortex being associated with enhanced stimulus detection (Sadaghiani, Hesselmann, & Kleinschmidt, 2009). A straightforward prediction from these studies would be that stimulus-induced activations should also be increased in individuals with AVHs, if there already is spontaneous hyper-excitation caused by the hallucinations, together with enhanced or better cognitive processing of an auditory task. However, the available empirical evidence seems to indicate the opposite: AVHs are associated with reduced neuronal activation to external stimulation in the auditory cortex ( David et al., 1996, Hugdahl et al., 2009 and Woodruff et al., 1997), and also reduced efficiency of speech processing in behavioural tasks ( Green et al., 1994 and Løberg et al., 2004). A similar paradox can also be observed in electrophysiological studies. Hallucinations have been found to result in reduced N1 amplitude to concurrently presented auditory stimuli ( Hubl, Koenig, Strik, Garcia, & Dierks, 2007), whereas ongoing hallucinations produce increased coherence measures across bilateral temporal electrode sites ( Sritharan et al., 2005). We propose that this discrepancy, with reduced processing capacity to exogenously presented auditory sounds visible in physiological measures and co-occurring with impaired behavioural performance, and enhanced activation in the same brain areas for endogenously induced “voices” is an apparent paradox in the existing literature. If auditory hallucinations are associated with neuronal hyper-excitation in the absence of an external stimulus, leading to spontaneously increased activation, one would then expect further increase in activation when an external auditory stimulus is presented. This paradox could suggest either a “refractory” effect in the auditory cortex, such that external stimuli are not capable of producing further activation sufficient to overcome the ongoing, spontaneous neural activity ( Hugdahl et al., 2009), or an aberrant attentional bias, such that the prefrontal top-down control mechanisms do not bias the auditory cortex sufficiently towards external stimuli. However, an un-systematic reading of the literature is not enough to establish whether the brain regions typically showing reduced activation during external auditory stimulation in patients who experience AVHs in fact are overlapping with the brain regions which have been found to increase in activation during AVHs in the absence of external stimulation because of small and differing sample sizes and variation in methods and data analysis ( Wager, Lindquist & Kaplan, 2007). For this reason we performed a systematic search through the literature, and conducted a meta-analysis, comparing neuroimaging studies of AVHs reporting endogenously driven (in the absence of an external auditory stimulus) versus exogenously driven (in the presence of an external auditory stimulus) activations. We have labelled the first kind of studies “endogenously evoked processing”, and the second kind of studies “exogenously evoked processing”. We expected to see converging effects of both increased and decreased activation in the same brain regions in patients who experience AVHs if our initial unsystematic observations were correct. Meta-analytical methods are gaining rapid popularity in the field of functional neuroimaging, providing a powerful tool for detecting consistent patterns across several different studies(Wager et al., 2007). While traditional meta-analyses are focused on the size of an effect across multiple studies in order to establish whether an effect is reliably present, in neuroimaging the interest is predominantly concentrated on the distribution of activation peaks in the brain volume (Wager, Lindquist, Nichols, Kober, & Van Snellenberg, 2009). In the present meta-analysis we aimed to systematically review available functional neuroimaging studies for the evidence that the paradox actually exists, and consequently the implications this may have for theories and models of the neuronal underpinnings and mechanisms behind AVHs. If neuronal activations are opposite in overlapping and converging brain areas, in particular in the auditory cortex to endogenously evoked versus exogenously evoked processing in AVH patients, this would have implications for interpretations of “state versus trait” effects of AVHs (cf. Kühn & Gallinat, 2011). Identifying a discrepant pattern of activation in endogenous and exogenous processing modes would also have implications for the understanding of how AVHs originate in the brain. We have restricted the search to functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) studies since these methods are based on hemodynamic changes related to neuronal activity. We employed the activation likelihood estimation (ALE) method for the meta-analysis, which allows to estimate the spatial convergence of reported activation peaks across multiple studies (Eickhoff et al., 2009).