Brain, Vol. 124, No. 11, 2335-2338,
November 2001
© 2001 Oxford University Press
Book reviews |
NEURONAL MECHANISMS OF MEMORY FORMATION: CONCEPTS OF LONG-TERM POTENTIATION AND BEYOND.
By Christian Hölscher. 2001. Cambridge: Cambridge University Press. Price £65. Pp. 528. ISBN 0-52177-067-X.
MRC Centre for Synaptic Plasticity, Department of Anatomy, University of Bristol, Bristol, UK
The driving force behind this book is the enduring question of whether synaptic plasticity is a suitable model for understanding the basis of learning and memory. In this book synaptic plasticity is taken almost exclusively to mean long-term potentiation (LTP) and there is virtually no discussion of, for example, long-term depression or the cerebellum.
The book begins with an introduction by Holscher in which some of the main arguments contained within the rest of the book are rehearsed. This begins with an assessment of the validity of in vitro slice techniques and then questions whether in vitro studies are relevant to learning; whether models of LTP are relevant to learning; and whether knockout techniques and the employment of these are useful for understanding firstly LTP and then learning? This is followed by a critique of the different approaches that have attempted to correlate LTP with the cellular processes that may be involved with learning. Most people in this field will recognize the problems and limitations of the systems that we work with, be these in vivo or in vitro. Most realize that it is not a simple matter to relate synaptic plasticity in a dish or in anaesthetized animals to learning in animals or in humans. Nevertheless, the Introduction is a readable account and provides a useful summary of some of the main issues.
The rest of the book is divided into five sections. The first of these deals with `Long-term potentiation in vitro and in vivo'. Abraham discusses the different types of LTP that are found in different brain regions, the types of stimuli usually employed to induce LTP and the underlying mechanisms of induction in different regions. There is an interesting section on neurogenesis (unique to dentate gyrus) and how this may influence plasticity and function of this region. The, not unreasonable, conclusion is reached that different forms of plasticity may subserve different functions in different brain regions.
Perhaps some of the best evidence to date linking synaptic plasticity and learning is from work in the amygdala. This is discussed in two related chapters by Rogan and colleagues and by Maren. First, fear conditioning is described before a summary is given of the classical work illustrating that increases in evoked potentials occur following pairing of an unconditioned acoustic stimulus with an unconditioned aversive stimulus. This is followed by a description of work in vitro investigating the role of protein kinase A in synaptic plasticity in the amygdala. The chapter by Maren describes how an NMDA (N-methyl-D-aspartate) receptor-dependent phenomenon may play a role in contextual fear conditioning. Whilst the role of hippocampal plasticity per se has not been proven, Maren describes various attempts that provide a correlation between this and learning. The final chapter in this section by Jeffery discusses experience-dependent plasticity of hippocampal place cells. It is strongly argued that this is a valid, and perhaps one of the best, model(s) for identifying a role for plasticity in hippocampus.
Section two deals with `synaptic plasticity induced by natural stimulation frequencies'. The first of these chapters by Pike and colleagues argues that the use of such patterns of stimulation in the in vitro hippocampus will provide a greater understanding of both plasticity and learning. The authors propose that bursts of activity that resemble hippocampal spiking can facilitate the induction of LTP. They propose a model in which inputs into the hippocampus are encoded by complex spikes and the recall of information occurs through simple spike-like activity. Maroun and co-workers discuss plasticity of local circuits in hippocampus and amygdala, and how changes in the activity of a circuit may be more important than the simple phenomenon of LTP per se. The plasticity of local circuits relies on decrements in GABA inhibition. They suggest a correlation exists between age-related spatial memory abilities and local circuit plasticity, but not with LTP. Holscher argues that theta frequency induction of LTP is a better model for memory formation than LTP induced by tetanic stimulation. Theta oscillations set up temporal constraints on the induction of LTP and allow the induction of LTP with much `weaker' stimulation. Finally, Munk addresses gamma oscillations and suggests that networks firing together at high frequencies may be able to process complicated information in a parallel manner. Furthermore, it is shown that gamma oscillations can facilitate the induction of synaptic plasticity.
Section three deals with `models from data of synaptic plasticity'. Lisman and colleagues discuss whether NMDA receptor-dependent LTP and oscillations in the hippocampus provide a mechanism for functions of this region. They argue that the patterns of gamma and theta oscillations can explain some of the properties and capacity of short-term memory. Rolls argues that LTP-like phenomena are most likely to play a role in learning and memory and discusses several lines of evidence which support the notion that Pavlovian conditioning relies on Hebbian plasticity within networks of neurones. In the very next chapter, however, McEachern and Shaw argue very strongly that LTP is not a good model for learning. They base their discussion on various arguments. For example, that LTP is not permanent, LTP is not input specific under all conditions and that LTP associativity is just as likely to play a role in pathology as in learning. Finally they argue that the LTP literature is so full of contradictions between different laboratories that this in itself is a weakness. McEachern and Shaw suggest that LTP may provide a better understanding of pathology than of learning. In the final chapter in this section Matzel and Shors argue against the associative LTP = associative learning hypothesis. They suggest that an increase in synaptic transmission, even within a network, simply cannot provide for the complexities of learning. In addition, they also agree with the McEachern and Shaw view that many of the properties of LTP are simply not sufficient to provide for associative learning.
In section four, `Setting the stage for memory formation', Cain discusses the limitations of one of the most popular learning tasks (the Morris water maze) in which correlations are often made between the pharmacological block of LTP and pharmacological block of learning. However, Cain argues that interpreting results from the water maze is fraught with difficulty and cites the effects that non-spatial pretraining has on the NMDA receptor-dependence of the task. Rose and Diamond make the argument that simply finding a correlation between lack of LTP and lack of some form of learning is no good reason to suppose that the latter requires the former. They cite contradictory examples from studies in aged animals to suggest that the correlation between LTP and learning is not always very strong. McNaughton in the next chapter also argues that LTP may not be a good model upon which to base a mechanism of learning and memory. First, he suggests that the hippocampus, where most of the plasticity work is conducted, is not the repository of memory and is therefore not the place to study. The hippocampus, he argues, plays a role in fear and anxiety but not memory per se. In the last chapter in this section Diamond and co-workers also suggest that the hippocampus does not play a role in learning and memory but is important in regulating stress. They argue that the amygdala and hippocampus can thus form a system in which the amygdala functions to produce emotional memory that is in some way operated on by the hippocampus. LTP may affect this system but is not directly involved in producing the conditions for this system.
Section five, `transgenic mice as tools', consists of three chapters on in vivo recording of place cells from transgenics, synaptic plasticity in genetically modified animals and what gene activation can tell us about synaptic plasticity. Cho and Eichenbaum discuss work on place cells and how the development of place fields is affected by various gene deletions. They suggest that plasticity mechanisms are involved in the organization and development of hippocampal place fields. Chapman provides a review of some of the genetic manipulations that have been used in the study of synaptic plasticity and learning and memory. He also provides a critique of common problems and difficulties of interpretation of the results from such studies. Inevitably, until more targeted (spatial and temporal) techniques of manipulation are used, the problems associated with traditional gene knockouts will continue to haunt the field. Davis and Laroche also point out the various pitfalls of global gene deletion but are optimistic that the new techniques will overcome previous problems.
Overall, this book provides a useful and critical analysis of some of the underlying assumptions that LTP provides a good model for understanding mechanisms of learning and memory.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||