Conditioned Blocking: Schizophrenia
377
et al, 1996) correlated negatively with the ratio of negative to positive symptoms in the
psychotic group (л=23, r= — .62, p < .002). Subgroup performance on these measures was
examined further.
An ANOVA with repeated measures for four CB measures [i.e. BA 1, BA 1-3, BA 1-5,
unblocking (F(9,141)=2.01, p = .04] gave a significant main effect for group [F(3,60) = 2.67,
p=.05] and an interaction [F(9,180)=2.3, p<.02; Figure 3]. The main effect showed an
average attenuation of — 2.5 s in the NP group vs an average CB effect in the other groups
of +1.3 to +2.4 s.
On the first trial-pair post-hoc tests confirmed that NP patients showed less CB than
either of the comparison groups and PH patients less than the controls. On BA 1-3 the NP
showed less CB than the controls and tended to show less than the PH group (Figure 3).
One-way analysis for unblocking [F(3,60) = 3.7,p<0.02] confirmed continuing less CB in
the NP than in the PH, OCD and control groups. Thus, the PH group showed only a
transient reduction of CB on the first trial-pair, but the NP group had a marked deficit
across measures. Even so, all groups demonstrated some learning about the added stimulus
across the test exposure. The following sections consider the associations of these CB
measures.
Personality Traits and Symptom States
Associations with positive and negative symptom clusters
Does a more consistent CB attenuation in NP than PH patients relate to a particular
symptom cluster? Both groups showed a similar incidence of the five negative symptom
clusters, but the PH group differed significantly on each of the four positive symptom
clusters [H[9,13) = 2.7, ρ = .05; Newman-Keuls, p = .05-.001 ].
In psychotic patients, reduced CB correlated with an increase of the ratio of negative to
positive symptoms (BA 1-3, unblocking, r= — .62 to —.67, ∕><.002; significant after a-
adjustment). Within the NP group the negative relationship was maintained (BA 1, BA 1-
3 and unblocking, r=.65 to —.89, ρ=.03-.0001, significant on а-adjustment) but an
opposite trend was seen in PH patients (r=+.6, p = .05). The apparent paradox that
negative symptoms were related to decreased CB in NP but to an increase in PH patients,
is explained by specific negative symptoms being relevant in the NP group (flat-affect vs
BA 1, r= —.62,p = .04; alogia vs unblocking, r = — .69, p = .02) but positive symptoms were
important in the PH group (delusions vs BA 1-3, r= —.55, p = .06; bizarre behaviour vs
unblocking or BA 1-5, r= — .59 to —.65, p=.04-.02, uncorrected levels of significance).
A linear regression analysis for the main CB measures vs the ratio of negative to positive
symptoms for psychotic patients confirmed an attenuation of CB with increased negative
symptoms (—.65 to — .89, p = .03-.0002). Analysis of separate symptom clusters was only
significant for the first trial-pair measure. Extremes of positive and negative symptoms may
interfere with normal CB (flat-affect — .90, bizarre behaviour — .66,/?= .007; hallucinations
+ .54, p = .035; thought disorder +.63,/?= .01).
Within the PH group linear regression confirmed the influence of summed positive
symptoms (BA 1-3, — .70,/?=.03). This was supported in the brief psychiatric ratings by a
relationship with positive symptoms, absent for negative, depressive or sum scores (— .63,