Dr Ramasubbu argues against our suggestion that psychological rather than neurological factors mostly account for post-stroke depression (PSD). His/her main argument is that our results could be due to a bias in selecting a control group of patients with endogenous depression. He/she remarks that it is not surprising that endogenous symptoms were more frequent in this control group, whereas affective symptoms related to brain damage prevailed in patients with major PSD. This argument could certainly be appropriate if the aim of our study consisted in matching the main symptoms of patients with major PSD and with major endogenous depression, but this was clearly not the scope of our paper. Patients with major endogenous depression were, in fact, the group against which we matched various subgroups of patients with major PSD observed at various time intervals from stroke. The scope of these comparisons consisted in assessing whether patients with endogenous depression are more similar to patients with major PSD observed immediately after the stroke than to those observed in more chronic post-stroke periods. A similar strategy had been used in previous studies of this field conducted by our research group, since the distinction between various forms of PSD plays a critical role in the model of PSD proposed by Robinson et al (see Reference Starkstein and RobinsonStarkstein & Robinson, 1989, for review). This model is, in fact, substantially based on the distinction between two forms of PSD: (a) the major form, due to a lesion encroaching upon the left frontal lobes and indistinguishable from the clinical and pathophysiological points of view from the major forms of endogenous depression; and (b) the minor form, considered as a form of reactive (dysthymic) depression and having no specific anatomical substrate. In a previous paper (Reference Gainotti, Azzoni and RazzanoGainotti et al, 1997) aiming to test this original version of the Robinson et al model, the symptomatological profiles of patients affected by endogenous depression were matched with those in three groups of stroke patients, on the basis of DSM-III-R criteria, as having no depression, minor PSD, or major PSD. The following predictions were made: (a) if major PSD is indistinguishable from endogenous depression, whereas minor PSD is a reactive form of depression, then the symptomatological profile of patients with a major form of PSD should be more similar to that of patients with endogenous depression than to that of patients with a form of minor PSD; (b) if, on the contrary, no qualitative difference exists between major and minor forms of PSD, then patients with major PSD should be more similar to those with a minor form of PSD than to those with a form of endogenous depression. Our results clearly supported the second prediction showing that a continuum exists between major and minor forms of PSD.

To account for these and other data also inconsistent with the Robinson et al model, Herrmann & Wallesch (Reference Herrmann and Wallesch1993) proposed a restricted version of the model which assumes that only the forms of major PSD observed in the acute post-stroke periods are very similar to endogenous depression, whereas those observed in more chronic stages must be considered as reactive forms, mainly due to psychosocial factors.

To test this new version of the Robinson et al model with a strategy similar to that used in our previous paper, we matched the symptomatological profiles of patients with major PSD observed at various time intervals from stroke with those of patients with endogenous depression. Since the profiles of patients with major PSD observed at various time intervals from stroke were very similar, and were very different from those in patients with endogenous depression, we concluded that even the restricted version of the Robinson et al model is inconsistent with our data.

We therefore think that no methodological defect exists in this or in our previous study and that our data allow us to conclude that no qualitative difference exists either between minor and major forms of PSD or between forms of major PSD that arise at various time intervals from stroke.

Let us pass now to our suggestion that psychological factors, rather than neurological factors, mostly account for PSD. This suggestion was mostly due to the distinction that we have more analytically described in our previous study (Reference Gainotti, Azzoni and RazzanoGainotti et al, 1997) between motivated and unmotivated affective patterns. The term ‘motivated reactions’ refers not only to the reactive symptoms of anxiety, emotionalism and catastrophic reactions, whose prevalence in patients with PSD could be due (as Dr Ramasubbu suggests) to a bias in the selection of the control group, but also to a qualitative analysis of the responses given by patients in sections of the Post-Stroke Depression Rating Scale (PSDRS) devised to evaluate ‘depressed mood’, ‘guilt feelings’ and ‘thoughts of death and/or suicide’. In these sections, patients were requested to qualify their response by saying whether their bad mood, guilt feelings and thoughts of death were related to their actual condition or were independent from it. Patients with major or minor forms of PSD usually attributed depression to the consequences of stroke, felt guilty because they considered their previous lifestyle as partly responsible for their actual disease, and had stroke-related death worries, whereas patients with endogenous depression attributed guilt feelings to their moral worthlessness and had active suicidal tendencies, rather than death worries. Furthermore, in both of our studies, diurnal mood variations were motivated (i.e. linked to situations stressing handicaps and disabilities) in patients with major PSD, but unmotivated (with a prevalence of depression in the early morning) in patients with endogenous depression.

In conclusion, even if we share with Dr Ramasubbu some doubts about the validity of dichotomous endogenous/reactive classifications, we would stress two points:

1. (a) our criticism was addressed to an influential model based on such dichotomy: and

2. (b) our tentative hypothesis that psychological factors play an important role in PSD seems, at least in part, justified.