I've looked everywhere for this answer but still can't find anything. If the mind and body are separate substances how can the mind interact with the body? What's the mechanism for the interaction?
The SEP discussion of interactive dualism notes that there is no problem with interaction in theory https://plato.stanford.edu/entries/dualism/#UniSubDua:
The simplest objection to interaction is that, in so far as mental properties, states or substances are of radically different kinds from each other, they lack that communality necessary for interaction. It is generally agreed that, in its most naive form, this objection to interactionism rests on a ‘billiard ball’ picture of causation: if all causation is by impact, how can the material and the immaterial impact upon each other? But if causation is either by a more ethereal force or energy or only a matter of constant conjunction, there would appear to be no problem in principle with the idea of interaction of mind and body.
Even if there is no objection in principle, there appears to be a conflict between interactionism and some basic principles of physical science. For example, if causal power was flowing in and out of the physical system, energy would not be conserved, and the conservation of energy is a fundamental scientific law. Various responses have been made to this. One suggestion is that it might be possible for mind to influence the distribution of energy, without altering its quantity. (See Averill and Keating 1981). Another response is to challenge the relevance of the conservation principle in this context. The conservation principle states that ‘in a causally isolated system the total amount of energy will remain constant’. Whereas ‘[t]he interactionist denies…that the human body is an isolated system’, so the principle is irrelevant (Larmer (1986), 282: this article presents a good brief survey of the options).
Robins Collins (2011) has claimed that the appeal to conservation by opponents of interactionism is something of a red herring because conservation principles are not ubiquitous in physics. He argues that energy is not conserved in general relativity, in quantum theory, or in the universe taken as a whole. Why then, should we insist on it in mind-brain interaction?
HOW interaction takes place, is speculative. It is an open question, subject to scientific investigation.
I reference my answer as to where interaction takes place, as each model has a different answer as to how interaction would operate:
Sir John Eccles postulated that the mind to body interaction took place in synapses, and the other direction took place through mind reading the state of a multiple neuron complex, in a digital on/off code. Synapses automatically pre-load themselves to fire, and all one needs is for one vesicle to open for its enzymes to catalyze the release of the other vesicles, and the synapse firing. Synapses then take very low energy input, before they have a dramatically larger response. Eccles assumed the soul to body interaction used quantum uncertainty to trigger the synapses -- and the knowledge of how to do this was some unconscious function of the soul. Eccles had a religious view, so presumably souls would be designed with this capability.
Sir Roger Penrose proposed that quantum uncertainty applies within microtubes within all neural cells, and the response of these tubules is indeterminate per heisenberg uncertainty principle, hence mind can do a zero energy input to the brain. Why or how our mind can do this -- is not clear. Penrose has a non=-religious view, and mind for him is an emergent phenomenon. I believe he would assume evolutionary principles would lead to minds learning the trick of influencing bodies.
Sir Richard Swinburne in Mind Brain, and Free Will, limits his inputs from mind to brain solely to adjusting neuron chemistry within the heisenberg limits, and further limits willing to marginal moral choices. With his religious starting point, I believe he presumes this is a pre-designed capability.
JP Moreland in The Soul, holds that minds are wholistic and individual, and willing directly leads to action of the brain. A wholistic/direct assumption avoids the where/how question. Moreland thinks that willing directly leads to action, and a mechanism is unnecessary.