When I look at something, when I talk to somebody, when I write a few sentences about "what I have changed my mind about and why", the neuronal network in my brain changes all the time and there are even structural changes in the brain. Why is it that these changes don't come to mind all the time but remain subthreshold? Certainly, if everything would come to mind what goes on in the brain, and if there would not be an efficient mechanism of informational garbage disposal, we would end up in mental chaos (which sometimes happens in unfortunate cases with neuronal dysfunctioning). It is only sometimes that certain events produce so much neuronal energy and catch so much attention that a conscious representation is made possible.
As most neuronal information processing remains in mental darkness, i.e. happens on an implicit level, it is in my view impossible to make a clear statement why somebody changed his or her mind about something. If somebody gives an explicit reason for having changed the mind about something, I am very suspicious. As "it thinks" all the time in my brain, and as these processes are beyond voluntary control, I am much less transparent to myself as I might want, and this is true for everybody. Thus, I cannot give a good reason why I changed my mind about a strong hypothesis or even belief or perhaps a prejudice in my scientific work which I had until several years ago.
A sentence of Ludwig Wittgenstein from his Tractatus Logico-Philosophicus (5.6) was like a dogma for me: "Die Grenzen meiner Sprache bedeuten die Grenzen meiner Welt. — The limits of my language signify the limits of my world " (my translation). Now I react to this sentence with an emphatic "No!".
As a neuroscientist I have to stay away from the language trap. In our research we are easily misguided by words. Without too much thinking we are referring to "consciousness", to "free will", to "thoughts", to "attention", to the "self", etc, and we give an ontological status to these terms. Some people even start to look at the potential site of consciousness or of free will in the brain, or some people ask the "what is ..." question that never can find an answer. The prototypical "what is ..." question was formulated 1600 years ago by Augustinus who said in the 11th book of his Confessions: "Quid est ergo tempus? Si nemo ex me quaerat scio, si quaerenti explicare velim nescio. — What is time? If nobody asks me, I know it, but if I have to explain it to somebody, I don't know it" (my translation).
Interestingly, Augustinus made a nice categorical mistake by referring to "knowing" at first on an implicit, and second on an explicit level. This categorical mistake is still with us when we ask questions like: "What is consciousness, free will,..."; one knows, but one does not. As neuroscientists we have to focus on processes in the brain which rarely or perhaps never map directly onto such terms as we use them. Complexity reduction in brains is necessary and it happens all the time, but the goal of this reductive process is not such terms, that might be useful for our communication, but efficient action. This is what I think today, but why I came to this conclusion I don't know; it was probably several reasons that finally resulted in a shift of mind. i.e. overcoming Wittgenstein's straitjacket.