I think there's a widespread confusion between two types of cause in psychology. One (cause1) has to do with precipitants, and is the sort we are interested in when we are trying to intervene in causal pathways. The other (cause2) has to do with character, with a pattern in which some particular example is situated, and is the sort we are interested in when we are trying to understand what it is that we are dealing with.
CBT models frequently muddle these together. Take the famous 'downward arrow' technique in CBT. The idea is that we keep on interrogating a negative thought with a question like "Well, what's so bad about that, what does it say about you?" and then we can get to the 'core beliefs'. These are presented as deep-rooted precipitants of lots of surface automatic thoughts, when they are better understood as the broadest possible meaning-organiser of the thought in question.
For both types of cause, we can anticipate that, on the whole, the causes function as sine qua non's. Yet here I think we need to be careful - and this is because the 'ceteris' in the ceteris paribus explication of the causal necessity refers in each case not to causes in general, nor to the other type of cause, but only to causes of the same form. That is to say: when I say that the cause2 of a certain behaviour is a certain mental state, I am saying that in the present context it is not similarly a function or cause2 of another mental state also present. (This is not to say that a certain behaviour cannot be multiply causally2 determined, but only to say that the logic of the causal2 explanations necessitates the explicit ruling in of such multiple determinants which otherwise are simply ruled out by the explanation in question's functioning as such.) And similarly for cause1: if I say that the man's arm is shaking because1 of Parkinsons, I am implying that it would not shake were it not for the Parkinsons. An arm may shake both because of Parkinsons and because of anxiety/depression - i.e. if either were singularly removed we would still expect a tremor - but that would need to be said when enumerating what the cause of something is. We can talk about what a cause of a behaviour is, meaning a contributing cause1, but this does not itself amount to a causal explanation of the behaviour unless it is wrapped up into a more complete story about what the cause/s is/are. Anyway, the limit of the sine qua non character of the causal story is - I want to suggest - that we can't rule that a particular behaviour or emotional state would not have been caused1 even if the cause2 we posit for it had not obtained.
The necessity of distinguishing causes is particularly potent in psychosomatic medicine. By treating a physical symptom as a communication we are, I believe, saying something about its character. We are not saying that it is caused1 by that which it communicates. But we can also, I believe, view physical symptoms as effects1 of causes1 which happen to be psychological. So: I have a bad shock, and my body goes into cortisol or adrenaline or what-have-you production mode. This then causes1 a physiological change in me, which may in turn cause1 a certain psychological state, etc. etc.
Psychotherapy typically works on the assumption that an investigation of meanings will result in a change of behaviour or emotion. In order for this to work we must presume that self-understanding involves a translation of an appreciation of causes2 into a reconfiguring of causes1. We guide our own activity according to our self-understanding, and our agency is that meeting point of the two causes 1 & 2. We come to realise that many of our behaviours and experiences do indeed conform to (are caused2 by) a certain pattern. This allows us to think of alternatives, and to understand that we do indeed have a mind (a perspective), and to guide our behaviour anew. The behaviour now gives rise to different experiences, and our experience of the world is altered. Yet this is only one possible mechanism for therapy. Another version is contained in the Kleinian idea that, in therapy, there is a direct communication from one unconscious to the other. I believe that this commuincation is of a cause1 form.
At its worst, psychotherapy either degenerates (psychodynamically) into a search for merely possible causes2 of behaviour and feeling, or (behaviourally) into a search for precipitants (causes1) of behaviours. In the former case the patient is understood, or at least potentially understood, but a fat lot of good it does them. In the latter case they are not even met as a person.