My first studies focused on the pharmacology of memory in rats and then on protein synthesis and memory in mice. What propelled me to human work was a 1971 publication on retrograde amnesia in memory-impaired patients. Retrograde amnesia refers to memory loss for information acquired before the onset of amnesia. The patients in that study had memory loss extending decades into the past, and its severity was similar at all past time periods. This result completely perplexed me. It did not fit what I had come to understand from the animal literature, where retrograde amnesia was typically limited and also temporally graded, affecting recent memory more than remote memory. Indeed, retrograde amnesia in animals was thought to provide evidence for memory consolidation: the idea that memory becomes more fixed and less vulnerable to disruption as time passes after learning.

I had been in graduate school at MIT at the time when the noted amnesic patient H.M. was being studied there. We were all familiar with his story and the neurosurgery that he had undergone to relieve severe epilepsy. So, it was not a big jump for me to begin studying memory and amnesia in humans, initially with psychiatric patients who were prescribed electroconvulsive therapy (ECT) for depression. (ECT was known to cause transient memory impairment). What was needed, I thought, was a test of past memory that could sample past time periods equivalently – that is, assess information from different time periods which had originally been learned to the same extent. We ended up with a memory test for the names of television programs that had broadcast for only one season during the past seventeen years. The popularity of the programs was similar across time periods. ECT resulted in temporally graded retrograde amnesia, sparing older memories but impairing memories acquired up to three years before treatment. This work provided the first evidence that memory consolidation in humans can occur across a lengthy time period (a few years) and that consolidation depends on reorganization within long-term memory. This led us to many more studies of retrograde amnesia, remote memory, and to what is referred to as the standard model of consolidation (not my term).