Memory Systems: What Amnesia Separates
Declarative memory for facts and episodes can be impaired while some skill learning survives. This dissociation supports partly separable memory systems. Systems consolidation proposes changes in the brain networks supporting memory over time. Whether remote recall still involves the hippocampus depends partly on the memory and the recall task.
What a learner can do afterwards
- Distinguish declarative from non-declarative memory using evidence from a described case
- Explain what systems consolidation claims and what evidence supports it
- Distinguish a synaptic correlate of behavioural memory from evidence that the synaptic change contributes causally
1 · Read
Declarative memory concerns facts and episodes accessible to conscious recall. Semantic memory concerns facts; episodic memory concerns experiences in a place and time. Procedural memory is a form of non-declarative memory expressed in learned skills. Describing how to ride a bicycle is declarative knowledge; keeping your balance through practice tests a skill. Speech alone does not define the boundary.
You have studied how brain circuits control movement. Now separate performing a skill from recalling its practice. In an illustrative tracing case, errors fall from 12 to 8 to 4 across three days, but the patient cannot recall earlier sessions. Similar dissociations after medial temporal damage support partly separable memory systems. They do not imply that every skill is spared.
Consolidation stabilizes memory after learning. Synaptic consolidation involves changes at connections. Systems consolidation concerns slower changes in interactions among brain regions. In the standard model, hippocampal and cortical interactions support new declarative memories. Retrieval can rely more on cortical networks over time. This is changing dependence, not a rule that every memory leaves the hippocampus.
A model of changing support over time, not a physical transfer of a complete memory between storage boxes.
Some people with hippocampal damage recall older facts better than recent ones. This time gradient supports systems consolidation if learning strength and recall demands are comparable. It does not prove that all old memories become hippocampus-independent. Detailed recall of remote personal episodes may still involve the hippocampus.
Training can improve recall while synaptic strength changes. This association alone does not show that the measured change caused the memory. Selective disruption or restoration, with sensory and motor controls, can test a causal contribution. Even then, one measured synapse does not describe the whole distributed memory trace.
A skill test also needs controls. Many tracing errors could reflect poor vision or impaired movement rather than impaired learning. Check those abilities and use a task with lower motor demands when needed. Compare learning across practice, rather than treating one poor performance as evidence that memory is lost.
Skill learning can survive impaired recall of practice. Consolidation changes memory support over time. Synaptic and behavioural evidence must be linked with care.
Lesson, questions and flashcards written for this dataset. Tracing values are invented teaching examples, not patient measurements.
Five original readings written for this dataset about memory systems, consolidation and the interpretation of learning experiments.
More to read
Performance and recall ask different questions
A keyboard lesson can leave several kinds of memory. You might remember the room and the teacher's instructions. Your fingers might also become more fluent at a practiced pattern. Recalling the lesson and performing the pattern are different tests.
Episodic declarative memory concerns the experience of that lesson. Procedural memory concerns the learned skill. In some cases of medial temporal amnesia, repeated practice improves performance even though the person cannot recall earlier sessions. That dissociation supports partly separable memory systems.
It does not establish that every skill survives every hippocampal injury. Tasks differ, and injuries can affect several structures. Nor does describing a technique aloud measure the skill itself. Knowledge of instructions and practiced performance can coexist.
Consolidation at connections and across networks
A memory can change after the learning session ends. Synaptic consolidation concerns cellular processes that help stabilize changes at connections. Systems consolidation concerns changes in interactions among brain regions over a longer period.
In the standard systems account, hippocampal and cortical interactions support new declarative memories. Later retrieval of some information can depend more on cortical networks. This is a proposal about changing support, not a complete memory moving between sealed containers.
The distinction matters when interpreting an experiment. A stronger synaptic response measures a change at one scale. A change in which regions are needed for recall addresses another. Evidence at either scale needs to be connected to the particular memory task.
Does memory age explain the difference?
Suppose an injury is followed by better recall of old facts than recently learned facts. This time gradient is consistent with changing support for memories over time.
Other differences could matter. The older facts may have received years of repetition. The recent facts may have been learned only once. Questions about one set may also be easier than questions about the other. Comparing learning strength and recall demands makes memory age more informative.
A gradient still does not establish that every remote memory is independent of the hippocampus. Detailed recollection of a personal episode may continue to involve it. Conclusions about older facts should not automatically be extended to every kind of autobiographical recall.
Testing a synaptic contribution
Better recall and stronger synaptic responses after training establish an association. They do not, by themselves, show that the measured synaptic change caused the improvement.
A causal experiment changes a candidate mechanism and tests what happens to memory. It also checks other abilities needed for the task. If an intervention prevents a person from seeing or moving normally, poor test performance could arise without a memory impairment. Disruption and restoration experiments are more informative when these alternatives are addressed.
Even evidence for a causal contribution does not describe the whole memory. A measured connection can be part of a wider trace. State which behavioural result changed and which mechanism was tested.
A tracing score needs controls
Tracing requires a participant to see the path, understand the task and control a hand. A high error count after injury could reflect difficulty with any of these demands.
Check vision and movement before interpreting the score as a learning measure. A comparison task requiring less movement can help. Repeated sessions reveal whether performance improves, which is different from measuring how well someone performs on the first attempt.
Ask about recall of the sessions separately. Improving performance alongside poor session recall supports a difference between skill learning and episodic recall. Poor performance on both measures needs further investigation; it does not establish that all learning has been lost.
2 · Watch
Take it off screen
Where it sits
Where this leads
Jobs that lean on this skill. Follow one to see everything it is built on.
8 questions wait behind this lesson, each with its answer explained. Every answer feeds the sky: stars light as they are learned, and dim when it is time to come back.