According to your example about the night sky, a claim is merely an observation if everyone will make the same claim in response to the same evidence. This seems very porous and weak, but following that logic, I would argue that given evidence that "the brain generates our experience of self by merging information about our body from multiple sources, including the eyes, the skin, the ears, and even one’s internal organs" it does not obviously follow that by observing a visualization of your body and heart beat will cause one to experience the phenomena of dissociation with his body. Then it seems that claiming so and contriving repeated samples is scientific.
It seems to me, the point you're really driving at is that a claim must be couched in some systemic view. It must base itself on a collection of related reinforcing ideas and assumptions (and state as much), and refine or extend that system. These systems abstract irrelevant things. Neuroscientific work is not phrased in string theory. It builds on a system that includes physical assumptions. Should a theoretical physicist argue neuroscience is on shaky ground because we do not understand all the details of the underlying physics? Should a neuroscientist argue evolutionary anthropology is on shaky ground because we don't understand all the details of the underlying neuroscience?
To your broader point. The motives and actions of neither psychiatric pharma, nor society, nor academia even, do not invalidate the possibility that collections of reinforcing ideas and assumptions can explain human thought at some level of neuroscientific abstraction. For better or worse, Jung, Freud, etc have established such systems which can and have been subject to methodical examination and testing by others. That they have subsequently not held up does not mean that the possibility is invalid.
I think also that there's a tendency to conflate the physicality (non-qualiality) of phenomena and its ability to be subject to "scientific" methods. And yet entrepreneurs expand knowledge with this all the time, as they make predictions about customer behavior, statistically test, refine, extend... This might not be Los Alamos "science", but it's method is certainly "scientific".