Ribot's Law is the surprisingly old theory that in memory loss due to
retrograde amnesia, the most recently formed memories will disappear the fastest. In other words, instead of the oldest memories fading the worst as happens normally,
amnesiacs' newest memories will be most likely to vanish. The law was first formulated in 1882 by French psychologist
Theodule-Armond Ribot, and he considered it a true law based on presumable brain
physiology rather than a theory based on available information. Since then forms of retrograde amnesia which do not follow the "law" (
Korsakoff's Syndrome springs to mind), so it should probably be referred to as
Ribot's Principle, although calling it a law is popular in the literature.
Implicit in Ribot's Law is the notion that the longer a memory exists in the brain's system, the more integrated it will be with other memories. Since it is more integrated in the system, its influences won't slacken even if a few neurons are killed or disconnected. Whereas memories of the long past only disappear due to consolidation, memories of the near past are immune to consolidation but can be dissociated entirely (or at least made difficult to retrieve) from the rest of the system due to trauma. Another reflection of Ribot's Law is seen in subjects with progressively impaired learning -- the later any ability is learned, the harder it will be to learn because the brain's networking ability will be less.
Two of the easiest places to observe Ribot's Law in action are in electroconvulsive therapy and closed head injuries, like football concussions. ECT has always been known to cause recent event memory loss -- psychoanalysts once thought ECT's effectiveness was due to sudden loss of traumatic memories that caused the depression. A 1977 study compared memory of material learned 10 minutes before ECT to memory learned 18 hours before, and found that the 10 minute memories disappeared at a statistically significant higher rate. Interestingly, after a closed head injury football players can often recall the whole play and the impact for a few minutes, after which it is lost for good. This makes it appear that short term memory isn't effected by the concussion, while transfer or indexing to long term memory is.