For most animals, whether it's molluscs or mice, practice makes perfect. Learning to ride a bike (if you're human) or to catch a grasshopper mid-jump (if you're a robber fly) might not work perfectly the first time around but with a little perseverance the goal (and in some cases the meal!) might just be reached. This motor learning is central to performing the complex sets of co-ordinated movements that make up animal behaviour, and the `memory' associated with motor learning seems to be stored in changes in the strength of connections between neurons. Although this idea appears to underlie motor learning in all animals, Kandel and co-workers now suggest that as well as knowing about connections between neurons it is also important to understand the intrinsic properties of the neurons themselves. Can these intrinsic properties affect animal's motor learning? To find the answer to this question we must look in an unlikely place: the heart.
At first, the heart may not appear to have much relevance to motor learning, but the electrical properties of neurons are not so different from those of the heart muscle cells (myocytes). In fact, many of the ion channels that allow ions in and out of heart myocytes are also found in neurons. Over 25 years ago, the hyperpolarization-activated, cyclic nucleotide-regulated nonselective cation (HCN) channels were discovered in heart myocytes, but more recent work has shown that they are also expressed in the mouse brain. One of these channels, HCN1, was expressed in one of the key structures involved in motor learning – the Purkinje cells of the cerebellum, suggesting a possible role in motor learning. But what are the effects of HCN1 on behaviours in which motor learning is essential?
Answering this question is no easy task. Kandel and co-workers not only made mutant mice with a deletion of the HCN1 channel but also targeted this deletion to different regions of the mouse brain. They then put the mutant and wild-type mice through a series of behavioural tests worthy of an assault course! These tests included swimming through a water maze and balancing on an accelerating rod that required complex repeated co-ordination of motor output as well as discrete behaviours such as an eye blink conditioning. When expression of the deleted HCN1 channel was targeted to the cerebellar Purkinje cells, the discrete eye blink conditioning was unaffected but the more complex behaviours, such as swimming and balancing, were impaired. Targeting the deleted HCN1 channels to the forebrain and not the cerebellum caused no impairment in these complex behaviours. This evidence suggests that HCN1 channels in the Purkinje cells play a vital role in motor learning.
Not content with their behavioural evidence for the role of HCN1 channels in motor learning, Kandel and co-workers extended the study to reveal the specific effects of these channels on the Purkinje cells themselves. Purkinje cells were injected with negative current pulses. Those cells lacking HCN1 channels took longer to recover to their previous activity level than their wild-type counterparts. This suggested that the HCN1 channels allow Purkinje cell activity to be independent of previous activity. Such a function would be critically important during complex repeated behaviours when Purkinje cells receive repetitive bursts of inputs from other neurons; without HCN1 channels, the Purkinje cells would still be recovering from one burst when the next one arrived.
This new study from Kandel and co-workers takes an unusual approach to motor learning in mice. It is likely that it will provoke many researchers to examine intrinsic neuronal properties and relate them to learning where previously they focussed on the relationship between neural connections and learning.
- © The Company of Biologists Limited 2004