نوروبیولوژی اسکیزوتایپی: اجسام مخطط سیگنال خطای پیش بینی در ارتباط با باورهای مانند توهم در افراد سالم

Healthy people sometimes report experiences and beliefs that are strikingly similar to the symptoms of psychosis in their bizarreness and the apparent lack of evidence supporting them. An important question is whether this represents merely a superficial resemblance or whether there is a genuine and deep similarity indicating, as some have suggested, a continuum between odd but healthy beliefs and the symptoms of psychotic illness. We sought to shed light on this question by determining whether the neural marker for prediction error - previously shown to be altered in early psychosis – is comparably altered in healthy individuals reporting schizotypal experiences and beliefs. We showed that non-clinical schizotypal experiences were significantly correlated with aberrant frontal and striatal prediction error signal. This correlation related to the distress associated with the beliefs. Given our previous observations that patients with first episode psychosis show altered neural responses to prediction error and that this alteration, in turn, relates to the severity of their delusional ideation, our results provide novel evidence in support of the view that schizotypy relates to psychosis at more than just a superficial descriptive level. However, the picture is a complex one in which the experiences, though associated with altered striatal responding, may provoke distress but may nonetheless be explained away, while an additional alteration in frontal cortical responding may allow the beliefs to become more delusion-like: intrusive and distressing.

Psychologically healthy people have an array of experiences, ideas and beliefs that may, on occasion, seem to overlap with those that characterize both emerging and established psychosis (David, 2010 and Kretschmer, 1925). Research that has focused on identifying and quantifying these beliefs and relating them to psychiatric illnesses has inspired the growing idea that psychotic experience exists as a part of a continuum (Linscott & van Os, 2010). Clearly, clinical psychosis entails experiences and beliefs that are sufficiently intrusive and distressing to have a marked effect on individuals' quality of life and functioning. However, it can be extremely difficult to develop an operational way of determining what is or is not a part of normal experience, and it is correspondingly difficult to distinguish clearly between a belief that is truly delusional and one that is merely unusual, arcane or irrational (Peters, Day, McKenna, & Orbach, 1999b). It is a challenge to develop our understanding of these healthy but strange experiences and beliefs and their implications for our understanding of psychotic mental illness. Do high scores on schizotypy scales, which document these experiences, reflect an increased vulnerability to psychotic illness? Does the existence of the symptom-like experiences in the healthy population prove that psychosis lies on a continuum with normal mental function? Establishing the phenomenological similarity between psychosis and schizotypy will only provide partial answers to these questions. To characterize the relationship more fully, we suggest that it is important to determine whether there is overlap at the neurobiological level. Our initial prediction, based on our own studies using purely behavioral measures (Corlett et al., 2009, Moore et al., 2011b and Teufel et al., 2010) was that our fMRI observations would favor a continuum model. However, key to this study was the acknowledgment that it is possible to have a behavioral similarity while, nevertheless, the underlying neural basis of altered beliefs in schizotypy and in psychosis might be quite different. In this respect, we see the fMRI measure as a valuable, even an essential tool, in addressing fully questions such as this. To that end, we sought to relate schizotypy to neural responses in an associative learning task. While undergoing functional magnetic resonance imaging (fMRI), healthy subjects completed a Kamin (1969) blocking task designed to reveal variations in patterns of prediction error (PE) signal across subjects. In blocking, prior learning leads to an attenuation of new learning such that there is a subsequently reduced expectation that a blocked stimulus has predictive power. Imagine that I have repeatedly learned that eating chicken causes an allergic reaction. If I eat a meal containing chicken and spinach, a subsequent allergic reaction is wholly predicted (by the presence of chicken in the meal) and I should develop no expectation that spinach has allergenic potential (in other words, learning about spinach has been “blocked”). If I now eat spinach alone and suffer an allergic reaction, this is relatively surprising: a prediction error results. This was the manipulation made in the current experiment. We made behavioral measures of expectancies as well as fMRI measures of neural responses during both blocking (low PE) and subsequent expectancy violation (high PE) trials in order to assay individual variability in blocking. Blocking was chosen because it enables a flexible characterization of PE signal (across low and high PE trials) and, moreover, it has already been explored behaviorally in healthy subjects whose responses are predictive of the severity of their attenuated psychosis-like experiences: for example, positive schizotypy scores on the OLIFE scale predict weaker blocking (Moore et al., 2011a and Moran et al., 2003) – consistent with these attenuated positive symptoms forming under the influence of an aberrant learning process, like clinical delusions (Corlett, Murray et al., 2007). Participants' schizotypal and related personality traits were quantified using the Chapman Scales (Eckblad & Chapman, 1983) and the Peters Delusion Inventory (PDI, Peters, Joseph, & Garety, 1999, see below). They then completed the blocking task during fMRI scanning. This entailed learning causal relationships between foods and allergic responses. In prior work, we found that inappropriate dorsolateral prefrontal PE signal during causal learning in patients with psychosis was predictive of the severity of delusions (Corlett, Murray et al., 2007). Evidence that aberrant right frontal PE signal relates to schizotypy would therefore favor a continuum model of psychosis ranging from high schizotypy in health to delusional belief in psychotic illness (Johns & van Os, 2001). However, a lack of comparable relationship between inappropriate prediction error responding and schizotypal features would, we argue, call into question any simple idea that schizotypy can be seen as an attenuated form of psychosis. Rather, such a negative finding would be more consistent with a conception of high schizotypy as a phenocopy of clinical symptoms: phenomenologically similar but neurobiologically separable and to some degree distinct from psychosis (Meehl, 1989).

Subjects PDI scores were comparable to published scores for healthy control subjects (Corlett et al., 2009 and Murray et al., 2010). Their mean number of endorsements on the PDI was five out of 21 questions (s.d.=3.1; patients with schizophrenia endorse 10.9 +/− 9.3, Corlett et al., 2009 and Murray et al., 2010). Our subjects' mean distress rating regarding their beliefs was 11.8 (s.d.=10.7), their mean pre-occupation with the unusual beliefs they endorsed was 11.9 (s.d.=9.6) and their mean conviction regarding those beliefs was 15.5 (s.d.=11.4). Again, these values are consistent with prior reports of PDI scores in healthy control subjects ( Corlett et al., 2009 and Murray et al., 2010).