Whether or not this is true, ECT memory loss is not hard to believe. I don't mean that the mechanism is obvious (it isn't), but only that the result is not very surprising.

Neurons communicate via electric signals. There are multiple reasonnable ways to implement a memory based on that substrate, but the interface is almost certainly going to be through axons.

There must be a way, whatever it may be, to query and to alter the brain's memory through its interface. We don't have to know precisely how the system works to predict that fuzzing the interface might result in adverse effects.

That remains true whether the underlying storage is in the field, in synapses, a more complicated change in gene expression, or some other property of the cell.