432
anomalously high (reflecting low activity) in post-
mortem samples from the left temporal lobe of
schizophrenics (Reynolds 1983).
Friston et al. (1992) also emphasized the frontal
connections in their rCBF analysis. It is in the dor-
solateral prefrontal and cingulate regions that
Petersen and colleagues (Corbetta et al. 1991)
found residual rCBF activation in healthy subjects
after subtracting brain images derived during pas-
sive focussed attention from those obtained under
conditions of divided attention. The cingulate is
also activated during performance of the incom-
patible condition of the Stroop test (Pardo et al.
1990: i.e. read the color of the ink of a word spell-
ing another color). Disorganized schizophrenics
more than others have problems on this part of the
Stroop test and it is this group which shows hypo-
perfusion of the cingulate (Liddle et al. 1992).
In the foregoing paragraphs strikingly mutually
supportive results on limbic function appear from
the comparison of modern imaging techniques
(MRI, PET and ERP), attempts to dissect atten-
tional mechanisms using tasks with specific de-
mands and use of schizophrenic subgroups defined
by symptom cluster rather than diagnosis.
I report briefly an exploratory analysis using this
type of approach. Two subtraction waves, mis-
match negativity (MMN) and difference negativity
(Nd) derived from the passive and active phases
of a З-tone oddball discrimination were compared.
The principle resembles Petersen’s subtraction of
PET-metabolic measures derived under different
stimulus-processing demands. The MMN is the
subtraction ofERPs elicited by common (0.8 KHz)
from rare deviant standard tones (2.0 KHz) under
passive non-attentional conditions. The mismatch
registered represents the difference in pitch and
frequency. The Nd is the result of subtracting ERPs
elicited by a rare 1.4 KHz tone in the passive
condition from the same tone used later as a target
for response.
The MMN is seen as representing a short-term
stimulus-trace-memory (Naataen 1990) reflecting
a largely automatic comparison process (pace Wol-
dorff et al. 1993). It was recently reported to be
markedly reduced in a heterogeneous group of
schizophrenics (Shelley et al. 1991). We confirm
this in Fig. 4 and show that the attenuation is
marked for both paranoid and non-paranoid groups
whether defined by clinical diagnosis or active
symptoms scored after the test in a semi-structured
interview. Further we show that there is a differ-
ence between sub-groups, but only if they are sepa-
rated by active symptoms. In the paranoid group
MMN is halved but this tends to be more than in
the non-paranoid group where MMN is >75% at-
tenuated at frontal sites. In the context of the em-
phasis above on reports of anomalous temporal
lobe function, it is of interest that particularly those
schizophrenics without paranoid symptoms show
an anomalous negativity at the T6 site (and less so
at T5: Fig. 4).
In contrast to the MMN, Nd is present in both
schizophrenic sub-groups (Oades et al. 1993). But
if the groups are separated by symptoms, Nd is
significantly attenuated over left frontal∕right
parieto-temporal sites in paranoid and over right
frontal/ left posterior parietal sites in the non-para-
noid group (Oades et al. 1994).
This result, obtained with teenaged patients soon
after admission, cannot exclude a potential influ-
ence of medication. The same subjects performed
a conditioned blocking task. This tests the extent
to which they learned about a superfluous stimu-
lus that was nonetheless relevant to task-solution.
While most paranoid patients performed relatively
normally, ignoring the redundant stimulus, non-
paranoids often showed attenuated blocking. How-
ever the non-ρaranoid group also showed increased
DA utilization, as measured in 24 h urine samples
Fig. 4. Top: Bilateral symmetric fronto-parietal distribution of MMN in healthy subjects (200-260 msec: see text) scale +3.0 to
-4.1 uv. Alongside t-test differences are compared to paranoid (P) and non-paranoid (NP) schizophrenics (line shows t(31) 1.7,
p < 0.05). Left: MMN in schizophrenics diagnosed as paranoid (12) and non-paranoid (16): A) MMN at 6 sites (bin 4 msec), B)
topography 200-260 msec (scale +2.4 to -2.5 uv), c) no significant t-test differences between psychotic groups (200-260 msec).
Right: MMN in schizophrenics with active paranoid symptoms (13) (rated post-test by median split of SANS-SAPS scores)
vs. non-paranoids (15). A) B) as above, C) t-tests between psychotic groups p < 0.05% at frontal sites (P > NP), T3 and T6
(NP > P).