This is an inference question. The passage suggests several ways in which dogs are superior to human methods of detecting seizures. We don't know how dogs can detect seizures, but we know some things about their behavior when they do detect seizures, and we know they can do so even when in another room. We also find out that epileptics do not display any number of "noticeable" changes before undergoing a seizure. Consider each choice, looking for a reasonable inference:

(A) The passage says nothing about accuracy of prediction, so this is not a reasonable inference.
(B) The passage is limited to information about seizures, so we know nothing about heart attack detection.
(C) This is one possible explanation of dogs' seizure detection abilities, but it is not supported by the passage, since there are other ways in which dogs might detect the onset of a seizure.
(D) This is correct. It is based directly on the information that dogs do not have to be in the same room as a patient. We don't know how they detect the onset of a seizure, but we know it isn't reliant on visual information.
(E) This choice reverses causality. The dog stares intently in response to the onset of a seizure; the seizure does not occur because the dog stares intently.