Some people say that free will is a myth because our actions are based on the sum of what are called "deterministic" brain activity. They are trying to say that chemical reactions always lead to the same result, so whatever we would do, we would ALWAYS do. Because it is determined by our body chemistry.
My problems with this "free will is a myth" viewpoint are many.
1. Those reactions are based on a Gaussian distribution of statistical results from a large number of small, individual reactions. If your body chemistry is "off" just a bit, those reactions can change. If your body's chemistry is affected by unknown external influences, those individual reactions might add up to something other than as expected.
2. We don't know the neurological pathways between external stimuli that require a complex response. So while we might talk about a reaction occurring, we don't know WHICH reaction is occurring and when. To say that it is deterministic is an oversimplification. Sure, it MIGHT be - but we don't know the mechanism.
3. Asimov's model of the positronic brain was based to a large degree on his knowledge of biochemistry (and he held an earned PhD in that field). He proposed that brains were like "difference engines" that tried to balance between the potential for one outcome vs. another. When the potentials were nearly balanced, you could have as many different outcomes as there were choices to balance. His model actually makes functional sense for the human brain. We see things in shades of gray and there is no telling just what we see at a given moment.
4. Having said all of the above, the pragmatic approach is that if we cannot reliably predict someone's actions because we don't know the mechanisms of decision-making, then for all intents and purposes, those decisions might as well have been made with free will no matter how deterministic the steps were that led to the decision.
Which brings to mind the Yale Law of Animal Behavior (quoted from Randall Garrett's "Unwise Child"). "Given two animals bred under rigidly controlled conditions for seventeen generations so that they are as close to identical as genetically possible, and then placed in identical situations, the two animals will do as they damned well please."