Lecture 3: Non-declarative memory

Procedural memory, such as the ‘know-how’ for performing an action, is part of non-declarative memory. Procedural memory is associated with the basal ganglia, which you may have already learned from neuroanatomy, a critical brain region for sensorimotor control. For sensorimotor control, the basal ganglia form circuitry with other brain regions, including the thalamus and the cortex. Lesions in the basal ganglia are commonly associated with Parkinson’s disease. Therefore, patients with Parkinson’s disease show deficit procedural learning. The performance of developing procedural skills can be demonstrated by the serial reaction time task. In this task, patients are asked to hit some buttons that represent the symbols shown here, with a fixed sequence. The sequence may consist of the same repeats or a random sequence.

For both healthy subjects and patients with Parkinson’s disease, it is difficult to learn to hit buttons properly with a random sequence. However, when it comes to the condition of repeated sequence, the patients with Parkinson’s disease will show a worse performance, compared to healthy subjects. The worse performance reflects that the patients may not well learn the skills to properly hit the button with the fixed order, i.e., a deficit in procedural memory.

In addition to the basal ganglia, the cerebellum also plays a key role in skill learning. But the cerebellum may play a more important role in developing and maintaining the automaticity of an action. What does that mean for ‘automaticity’? It would be interesting to know that most of our actions, even some ‘professional behaviour’, are done automatically, which means we can perform this action naturally, without any extra effort. For example, we can chew and swallow without noticing it. A dentist with higher proficiency in tooth preparation can perform all the steps even with a slight distraction from others. In other words, the brain will keep on running the action, even if we do not fully pay attention to it.

Except for sensorimotor learning, the acquisition of probabilistic association between things is also part of non-declarative memory. Here ‘probabilistic association’ means the association between things are uncertain. For example, when we predict weather from the sun and cloud, we know it may or may not rain – but we are not very sure if it must or must not rain. The probability of this prediction is sometimes implicit; it means we feel it difficult to explain why it may be rainy or sunny verbally. Learning of the probabilistic association is also associated with the basal ganglia. For example, patients with Parkinson’s disease will show worse performance in a test of probabilistic learning, compared to healthy subjects.

Finally, you may feel surprised that even memory can be even shaped by perceptual processing. This can be found in the priming effect, i.e., a change in a person’s ability to identify, produce or classify an item, as a result of a previous encounter with that item or a related item. In other words, we feel familiar with the things that we have found previously, especially when they are in common. For example, if I saw a red ball previously, I would pick up a red apple later in the market, because my mind has been primed by an image of ‘a red ball’.

This is a case of perceptual priming; it means we are influenced by something we met previously for similar colour, shape or size. There is also semantic memory, i.e., we are primed by the things with similar ideas. For example, in the following list, please identify the words related to a person. Many experiments showed that subjects responded to the word ‘driver’ and ‘sailor’ more quickly, compared to ‘doctor’ and ‘politician’. This is because the words right before ‘driver’ is ‘car’ and the word right before ‘sailor’ is ‘boat’ – we feel easier to spot on driver and sailor, because we have been primed by the word bus and boat, which bring the similar meaning with driver and sailor, respectively.

I know it is quite bizarre for some people when I say the learning of sensorimotor skills is also part of memory. Intuitively, if we are so familiar with doing something, that would become part of our habits. Yeah, our habits are also part of our non-declarative memory. They are ‘memorized’ in a non-verbal way, but they can still be assessed. For example, how do I know that you really master a sensorimotor skill, as such, well, playing an action computer game? What I would do is try to distract you during the play. The masters of playing an action game can still beat the highest score even when they are chatting with friends.

When we master the skill, the brain forms a network that mandates the sensorimotor processing automatically, and usually out of our conscious mind. Same to our habits – they are conducted automatically and mindlessly because they have been part of our memory.