It's not valid as a deductive argument, but it is Bayesian evidence

The "not a logical disproof" defense can be used against more or less all arguments drawing on empirical science. It would be more charitable to assume Richard means something like "the probability of brain damage by region delicately affecting particular parts of information processing is reasonably high given belief in souls." Although that still looks quite wrong. The "radio" analogy would not predict a speech center, or other patterns of brain damage and impairment that look like interfering with computation, not transmission.