Objectives. The major question examined in this paper is whether selec
tive attentional and interpretative processing of emotional informatio
n occurs in depression, and if so, whether it depends on a close match
between the material used and current concerns. Design. Twenty-four d
epressed patients and the same number of matched controls were tested
using two selective processing tasks (described below), and their perf
ormance related to self-reported sociotropic and autonomy-related conc
erns. Methods. Colour-naming interference and interpretation of ambigu
ous situations were assessed using material judged relevant to each of
the Sociotropy-autonomy Scales. Results. Depression was associated wi
th a general interference effect for all negative concern words, and m
ore negative interpretations of ambiguous situations, while controls s
howed a converse bias in favour of all positive interpretations. There
was no convincing evidence that this negative processing bias was pro
portional to the match between material and self-reported sociotropic
or autonomous concerns. Conclusions. Depressed patients showed evidenc
e of cognitive biases favouring all negative self-related information,
on both attentional and interpretative tasks. We suggest that such ef
fects in depression may occur only under conditions allowing the elabo
rative processing of negative material related to oneself.