Biological Sciences, Open University
RESEARCH GROUPS Behavioural and Biochemical Correlates of Learning and Memory

 God’s Organism? The chick as a model system for memory studies

Steven P R Rose

Brain and Behaviour Research Group,

The Open University, Milton Keynes MK7 6AA, UK

Learn. Mem. 2000 7: 1-17.
With thanks to Learning and Memory for permission to reproduce this review.





The young chick is a powerful model system in which to study the biochemical and morphological processes underlying memory formation. Training chicks on a one trial passive avoidance task results in a molecular cascade in a specific brain region, the intermediate medial hyperstriatum ventrale. This cascades is initiated by glutamate release and engages a series of synaptic transients including increased calcium flux, upregulation of NMDA-glutamate receptors, membrane protein phosphorylations and the retrograde messenger NO. Expression of immediate early genes c-fos and c-jun precedes the synthesis, glycosylation and redistribution, 4+ hours downstream, of a number of synaptic membrane proteins, notably NCAM and L1. Other membrane proteins required in the early phase of memory formation include the amyloid precursor protein APP and Apolipoprotein E. There are concomitant increases in dendritic spine number and changes in synaptic structure. Non-synaptic factors including corticosterone and BDNF can modulate retention of the avoidance response, enhancing the salience of otherwise weakly retained memory. These results are discussed in relation to general concepts of memory formation and the spatio-temporal distribution of the putative memory trace.

  Early learning in the chick


The distinguished biochemist Hans Krebs once remarked that God had provided just the right organism with which to study any particular biological problem. The purpose of this review is to argue that the young domestic chick rates high amongst God’s organisms for the study of the molecular and cellular mechanisms of memory formation. Chicks are precocial birds, and need actively to explore and learn about their environment from the moment they hatch. Thus they learn very rapidly to identify their mother on the basis of visual, olfactory and auditory cues (imprinting), to distinguish edible from inedible or distasteful food, and to navigate complex routes. Training paradigms that exploit these species-specific tasks work with the grain of the animal’s biology, and because such learning is a significant event in the young chick’s life the experiences involved may be expected to result in readily measurable brain changes. Chicks have large and well-developed brains and soft unossified skulls, making localised cerebral injections of drugs easy without the use of implanted cannulae or anaesthesia. The virtual lack of any blood-brain barrier in these young animals also ensures rapid entry into the brain of peripherally injected agents. Against these advantages must be weighed the relatively limited knowledge of the intimate neuroanatomy of the avian brain (but see Csillag, 1999), a degree of genetic heterogeneity, and the complication resulting from the fact that in such young animals task-induced changes in cerebral biochemistry, electrophysiology or morphology are superimposed on rapidly developing neural systems (Andrew, 1991; Rogers, 1995).

  Learning Paradigms


The study of both the phenomenology and the neurobiology of learning and memory in avian species has a long and distinguished history. Pigeons vied with rats as the favourite laboratory animals for behaviourists studying conditioning, but more ‘natural’ forms of learning have long fascinated. Marler, Konishi and Notttebohm pioneered the behavioural and neurobiological analysis of song learning in song sparrows, canaries and zebra finches (Marler, 1991;Konishi, 1989;Nottebohm, 1970; 1980), tracing neural pathways, identifying issues of lateralisation, hormonal and related effects which have found their parallels in work with the chick. Krebs, Sherry and their colleagues explored the role of the hippocampus in food-storing birds such as marsh tits and chickadees (Krebs, Sherry, Healy, Perry and Vaccarino, 1989). Some of this work has more recently begun to move to the biochemical and molecular level that I focus on here in the context of the chick. The prospect of using the young chick for the study of cellular and molecular memory mechanisms emerged in the 1960s. Bateson and his colleagues opted for the study of visual imprinting (for review see Horn, 1985). Imprinting, as a form of early learning in some avian species, had been known folklorically for many years (Bateson, 1966), but was drawn to scientific attention by the work of Lorenz (eg Lorenz, 1935) although his interpretations of the phenomenon were not uncritically received. The study of the neurobiology of imprinting has continued in Horn’s and other labs (McCabe and Nicol, 1999; Johnston and Rogers, 1998). Scheich, Braun and their colleagues explored auditory imprinting (Scheich, 1987; Braun, Bock, Metzger, Jiang and Schnabel, 1999). In these models, chicks are exposed for up to two hours to visual or auditory cues that come to represent ‘mother’ (flashing light, stuffed hen, tone). They will subsequently approach the familiar object when offered a choice between it and an unfamiliar stimulus.

Cherkin (1969) developed a one-trial passive avoidance learning task based on the young chick’s tendency to peck at small bright objects such as beads, discussed in more detail below. In an appetitive visual discrimination task the chick learns to distinguish food crumbs scattered on the floor of its pen from rather similar pebbles or beads glued to the surface (Andrew and Rogers, 1972; Tiunova, Anokhin and Rose, 1998). Isolated chicks will quickly learn a simple maze in order to rejoin their hatchmates (Gilbert, Patterson and Rose, 1989; Regolin and Rose, 1999). Barber et al (Barber A, Gilbert and Rose, 1989; Barber TA, Howorth, Klunk and Cho, 1999) have employed a sickness-aversion protocol in which chicks, having pecked a neutral object (a dry coloured bead) and subsequently been made sick by intraperitoneal injection of lithium chloride, will later avoid the bead. Social learning, in which one chick observes and learns from the experience of a second is another important paradigm, bypassing as it does direct physiological or pharmacological manipulation of the learner (Johnston, Burne and Rose, 1999). Bradley and his colleagues (Bradley, Burns, King and Webb, 1999) have explored forms of LTP in the chick forebrain which correlate with the training protocols described below. However this review is largely confined to the data from the passive avoidance model.

The passive avoidance task has the merits of being rapid and sharply timed (chicks peck a bead within 10 seconds) and as many as 60 chicks can be trained in a single session. In the standard version of the task in our lab, day-old chicks are held in pairs in small pens, pretrained by being offered a small dry white bead, and those that peck trained with a larger (4mm dia) chrome or coloured bead coated with the distasteful methylanthranilate (MeA) (Lossner and Rose, 1983). Chicks that peck such a bead show a disgust reaction (backing away, shaking their heads and wiping their bills) and will avoid a similar but dry bead for at least 48hr subsequently. However, they continue to discriminate, as shown by pecking at control beads of other colours. Chicks trained on the bitter bead are matched with controls which have pecked at a water-coated or dry bead, and which peck the dry bead on test. Generally some 80% of chicks in any hatch group can be successfully trained and tested on this protocol. Each chick is usually trained and tested only once. These protocols, coupled with close observation of the chicks, enables one to distinguish the general effects of the training experience with its accompanying visual, motor and gustatory engagements, from the cellular processes involved necessarily and specifically in memory formation for the task (eg Rose and Harding, 1984). This distinction is reflected in the data presented below. Other labs (see Andrew, 1991 for review) slightly vary the protocols in ways which may affect the temporal sequence of post-training behavioural and molecular processes (Burne and Rose, 1997). Nonetheless, where data is comparable, there is broad consensus concerning both the behavioural and biochemical sequelae of training amongst the several chick labs.

Both correlative and interventive strategies are possible. In the correlative approach, appropriate brain regions can be dissected from trained and control birds at specific post-training times and tissue processed for biochemical, immunocytochemical, autoradiographic or microscopic analysis. It is also possible to prepare tissue slices (prisms) or synaptoneurosomes from the chicks, enabling calcium and neurotransmitter fluxes to be determined. Braun’s group (Gruss and Braun, 1996; Daisley, Gruss, Rose and Braun, 1998) has developed a microdialysis system enabling such fluxes to be measured in vivo during both training and testing. In the interventive approach all the birds are trained on MeA, and injected either before or after training with drug, antibody or antisense, or vehicle, to explore the possible enhancing or amnestic effect of the agent. Chicks that peck the previously distasteful bead on test are considered to be amnesic for the training. As this pecking response requires a positive and accurate act by the bird, it also controls for effects of the agent on attentional, visual and motor processes.

The sharply timed nature of the learning experience, together with a combination of these experimental strategies, has enabled us to identify a biochemical cascade associated with memory consolidation in the minutes to hours following training. Thus a change in some biochemical marker at a specific post-training time, occurring in trained compared with control chicks, might imply its direct engagement in memory expression at that time. Alternatively it could indicate the mobilisation of that marker as part of a sequence leading to the synthesis of a molecule, or cellular reorganisation, required for expression of memory. A similar argument applies to the timing of the onset of amnesia following intracerebral drug injection.

The temporal cascade


2-deoxyglucose studies revealed that two regions of the chick forebrain show enhanced metabolic activity in the forty minutes following training (Kossut and Rose 1984). These are the intermediate medial hyperstriatum ventrale (IMHV), an association ‘cortical’ area previously implicated in visual imprinting (Horn, 1985), and the lobus parolfactorius, a basal ganglia homologue (Csillag, 1999). The chick brain is strongly lateralised (Andrew, 1999: Rogers and Deng, 1999) and many, though not all of the molecular events we have observed are confined to the left IMHV. It is this cascade, primarily as identified by our lab, which forms the focus of the present review, although this means that it is not possible to do full justice to the work of the La Trobe group of Ng and his colleagues which most closely parallels our own (for reviews see Gibbs and Ng, 1977; Ng and Gibbs, 1991; Gibbs, Ng and Crowe, 1991 and other references in this text, passim). The combination of correlative and interventive methods has enabled an approximate temporal sequence to be traced. Table 1 shows the effects of injecting specific interventive agents just prior to or just following training and the time of onset of amnesia consequent on that injection. The cartoon of Fig 1 interprets these and the correlative and enhancement data discussed below in terms of a temporal sequence of molecular events in the IMHV following the bead peck.

Time of onset of amnesia for the passive avoidance task
following injection of blocking agents just before or after training.
Process Blocked
Onset of Amnesia
1) Muscimol
GABA agonist
2) Nitroarginine
NO synthesis
3) TFP; W13; A3;
Ca2+/CAM kinase/B50?
4) w-conotoxin GVIA
N-type Ca2+ channels
5) 2-D-galactose*
6) Anti-L1*
Signal transduction
7) H7; HA156;
   H8; H9; ML9
neurogranin phosphorylation
8) c-fos antisense
c-fos expression
9) b-APP antisense
b-APP expression
10) Anisomycin*
Protein synthesis
11) NDGA
Arachidonic acid synthesis
12) AP5; MK-801
NMDA receptors
13) Dantrolene
Intracellular Ca2+ stores
14) RU 38486
Glucocorticoid receptors
15) anti-BDNF

*Anisomycin is also amnestic if injected 4hr post-training; anti-L1 and 2-D-gal if injected at 5.5 hr post-training. Agents only effective during the ‘second wave’ discussed in the text are not listed here.

References to Table 1

  1. Clements and Bourne, 1996
  2. Holscher and Rose, 1993
  3. Serrano et al 1994
  4. Clements, Rose and Tiunova, 1995
  5. Crowe et al, 1994
  6. Scholey et al, 1995
  7. Serrano et al, 1994
  8. Mileusnic et al, 1996
  9. Mileusnic et al, 1999
  10. Freeman and Rose, 1995
  11. Holscher and Rose, 1994
  12. Rickard, Poot, Gibbs, and Ng,1994
  13. Salinska et al, 1999b
  14. Sandi and Rose, 1994b
  15. Johnston et al, 1999

Pre/post Synaptic Parallel Processing in Memory Formation


The first hour


During training, and in the five minutes which follow, there is enhanced release of glutamate in the IMHV, detected both by microdialysis in vivo (Daisley et al, 1999), and in tissue slices (Daisley and Rose, 1994). Over the same time period there is also an increase in potassium -stimulated calcium concentration in synaptoneurosomes isolated from the IMHV (Salinska, Chaudhury, Bourne and Rose, 1999). Within the succeeding forty minutes, although we cannot assign them a precise temporal dependency, we have found: increases in NMDA-stimulated calcium flux in synaptoneurosomes (Salinska et al, 1999); in ligand binding to the NMDA-glutamate receptor (Stewart, Bourne and Steele, 1992; Steele, Stewart and Rose, 1995) and of phosphorylation of the presynaptic membrane protein B50 (aka GAP43), (Ali, Bullock and Rose, 1988) coupled with a translocation of cytosolic PKC to the membrane (Burchuladze, Potter and Rose, 1990). There is increased release of the putative retrograde messenger arachidonic acid, in tissue prisms prepared 30-75min post-training, though the onset time for amnesia if the arachidonic acid synthesis is blocked with phospholipase A2 inhibitors is delayed until 75 minutes (Clements and Rose, 1996; Holscher and Rose, 1994). Interventive studies with MK801 (Burchuladze and Rose, 1992), the N-type calcium channel blocker W-conotoxin GVIA (Clements, Rose and Tiunova, 1995) and PKC inhibitors such as mellitin and H7 (Burchuladze et al, 1990), injected into the IMHV either just before or just after training, all produce amnesia with an onset time of 30min to an hour. GABAA agonists are also amnestic at this time (Clements and Bourne, 1996); so too is nitroarginine, which blocks synthesis of the putative retrograde messenger NO (Holscher and Rose, 1993; Rickard, Ng and Gibbs, 1998). Other labs have found an involvement of a variety of protein kinases, notably PKA, over this period (Serrano, Rodriguez, Bennett and Rosenzweig, 1995).

Most of these processes have been observed in other learning paradigms in mammals, and are also reminiscent of the sequence identified in the induction phase of NMDA-dependent LTP (Izquierdo and Medina, 1997). (An exception is our failure to observe any effect on retention of blocking post-synaptic L-type calcium channels with nimodipine. The N-type calcium channels blocked by conotoxin are generally assumed to be presynaptic in mammals, although this needs to be confirmed in the chick.) However, the merits of the brevity of the training experience in the chick and the fairly sharply timed onset of amnesia following various pharmacological interventions do make it possible to propose a rather more precise sequence for the various steps in the cascade than is otherwise feasible. As indicated in Fig 1 the observations also cast some light on the long-running discussion of pre versus post-synaptic events in the initiation of LTP (Bliss and Collingridge, 1993). Thus NMDA upregulation is temporally downstream of the initial in vivo increases in glutamate release and in vitro increases in potassium stimulated calcium flux. These early events, together with the changes in phosphorylation of the presynaptic B50/GAP 43, and, if taken at face value, the conotoxin data, all point to presynaptic engagement. The amnesia resulting from blocking NOS with nitroarginine, which sets in by thirty minutes, implies two-way signalling between post and pre-synaptic sides, and that this continued traffic is necessary if both are to undergo reconstruction during the later stages of memory formation.

Thus it would appear that the training experience generates a sequence of rapid synaptic transients which provide a temporary ‘hold’ for the memory — the phases categorised as short and intermediate term memory by Gibbs and Ng (1977; see also Patterson, Alvarado, Rosenzweig and Bennett, 1988). As well as forming the brain substrate of the remembered avoidance over this period, these transients must serve two other functions. They must initiate the sequence of pre- and post-synaptic intracellular processes which will in due course result in the lasting 'synaptic' changes presumed to underlie long term memory, and they must also serve to ‘tag’ relevant active synapses, perhaps via B50 and other membrane phosphorylations so as to indicate those synapses later to be more lastingly modified.

1-8 hours


A key step in the intracellular cascade must be the link between synapse and nucleus. Calcium is clearly a major player here, and that intracellular calcium signalling may be important is indicated by our recent observation that within 10 min post-training there is also a mobilisation of synaptoneurosomal ryanodine-sensitive calcium stores (Salinska, Bourne and Rose, 1999), whilst dantrolene, which blocks calcium release from these stores, injected 30 min pre- or 30 min post-training, produces amnesia by 3hr post-training. Synaptoneurosomes are largely presynaptic, though they contain resealed postsynaptic (dendritic) elements as well, so it is not possible to distinguish whether the mobilised calcium stores are located at one, the other or both sides of the cleft.

That activation of a number of transcription factors must be amongst the next steps in the process is clear from the elucidation of a role for CREB in several mammalian learning paradigms.We however have focused on the role of immediate early genes, c-fos and c-jun. That visual stimulation alone is sufficient to activate fos or jun expression is well known, as it was not surprising that we found that passive avoidance training elicited such activation (Anokhin, Mileusnic, Shamakhina and Rose, 1991). To control for the specificity of the expression we compared overtrained and undertrained chicks on the pebble floor task described above. Although both groups showed the pecking behaviour, the learning, as opposed to the already trained group, showed a clear increase in jun expression (Anokhin and Rose, 1991).Further evidence as to the necessity of fos expression for longer term memory is provided by the observation that antisense to c-fos, given 6 or more hours pretraining, blocks its synthesis (Mileusnic, Anokhin and Rose, 1996) and chicks become amnesic within 3hr post-training.

One of the few universal findings in studies of biochemical processes in memory formation is that long(er) -term memory is protein synthesis dependent (Davies and Squire, 1984). Passive avoidance training is no exception, and anisomycin injected into the IMHV either before or up to some 60 minutes post-training results in amnesia for the avoidance. If the anisomycin is injected before training, amnesia sets in by the end of the first post-training hour, leading to the suggestion that beyond this period memory is protein-synthesis independent (Gibbs and Ng, 1977). However, the earlier view that beyond this time a protein synthesis-independent, long-term memory has been established is no longer tenable. Whilst anisomycin injections 2 and 3 hr post-training are without effect on memory, injections given 4 or 5 hrs post-training are amnestic in animals tested at 24 hr (Freeman and Rose, 1995) as earlier found for rats by Grecksch and Matthies (1980). Thus there is a second, downstream wave of training-related protein synthesis that we interpret as being the period during which late genes are activated and structural proteins synthesised. However, this is not the entire explanation; other processes are also mobilised during this later time period. Whereas 30 minutes post-training NMDA but not AMPA receptors are engaged, 5.5 hr post-training AMPA but not NMDA receptors are upregulated, leading to an increase in AMPA-stimulated calcium flux in synaptoneurosomes. CNQX given at this time is amnestic (Steele and Stewart, 1995; Salinska et al, 1999). The relationship between these events and the protein synthetic wave is still not clear.

Whilst much attention within the learning and memory community is currently directed towards the roles of the many transcription factors involved in the early phases of memory formation, we have focussed on identifying the later gene products. There is considerable evidence for a role for synaptic membrane glycoproteins (Popov, Schulzeck, Pohle, and Matthies,1981, for reviews see Matthies, 1989; Rose, 1995a,b). Amongst the membrane glycoproteins are a variety of receptors, and the family of cell and matrix adherence molecules. The cell adhesion molecules (CAMs) are transmembrane molecules, whose glycosylated extracellular domains can bind either homophilically or heterophilically, providing a mechanism for associating pre- and post-synaptic membranes.Their potential role in synaptic plasticity has long been emphasised by Edelman (1985) (see also Pollerberg, Sadoul, Goridis and Schachner, 1985; Goelet, Castelluci, Schachner and Kandel, 1986). NCAM in particular exists in a number of isoforms. In its immature form is relatively highly sialylated, and less adherent. The mature form is desialylated, and more adherent. This has led to the suggestion, for which there is now good empirical evidence (Doyle, Nolan, Bell, and Regan, 1992a) that synaptic restructuring, either in the context of general development, or more specifically as a result of training experiences, requires first a sialylating, deadherence step, enabling synaptic mobility to occur, and subsequently desialylation and readherence. However these are not the only roles proposed for the CAMs; they are also likely to be involved in transmembrane signalling.

Fucose is a general glycoprotein precursor, and training chicks on MeA results in a long-lasting increased fucosylation of synaptic membrane glycoproteins. The use of the competitive inhibitor of galactose incorporation into glycoproteins, 2-deoxygalactose (2-d-gal), which prevents terminal fucosylation was pioneered by Matthies and his colleagues. Injected into the IMHV either around the time of training, or 6-8hr later, 2-d-gal results in amnesia in chicks tested at 24hr (Rose and Jork, 1987; Scholey, Rose, Zamani, Bock, and Schachner, 1993). This suggests a double wave of glycoprotein synthesis analogous to that already found with anisomycin, and comparable to that observed by the Matthies group following a brightness discrimination task in rats. These time windows are displaced with respect to the effects of anisomycin, occurring prior to anisomycin-induced amnesia at the first time point, and after it at the second (Freeman and Rose, 1995; Crowe, Zhao, Sedman and Ng, 1994). We interpret this as indicating that during the first period 2-d-gal is blocking post-translational glycosylation of already synthesized proteins, whilst at the second time period it affects de novo glycoprotein synthesis.

Among the glycoproteins whose synthesis is blocked by 2-d-gal are a number of CAMs. Two in particular are required for longer term memory; NCAM and NgCAM (aka L1) (Scholey et al,1993; Scholey, Mileusnic, Schachner, and Rose,1995). Both have also been shown to be required for maintenance of LTP (Luthi, Laurent, Figurov, Muller, and Schachner,1994). NCAM levels also increase after visual imprinting (Solomonia, McCabe and Horn, 1995). Training chicks on MeA results in a redistribution of NCAM within the synaptic junction, as detected by immunogold labelling (Rusakov, Davies, Krivko, Stewart and Schachner, 1994; Skibo, Davies, Rusakov, Stewart and Schachner, 1993). Specific blocking of NCAM synthesis with antisense, injected over the 24hr post-hatch period before the birds are trained, does not prevent the chicks learning the avoidance, but amnesia sets in within 3hr (Mileusnic, Lancashire and Rose, 1999). However, interference with the functioning of already synthesised CAM molecules is also amnestic. Thus if antibodies which bind to the extracellular domains of either NCAM or L1 are injected into the IMHV at 5-6hr post-training, chicks show amnesia when tested at 24hr (Scholey et al, 1993; 1995), a time at which the antibodies themselves are no longer detectable in the brain. Antibodies to NCAM are not amnestic if injected at other times, but antibodies to L1, injected 30 min pretraining, are also amnestic when the chicks are tested at 24hr. The extracellular domains of L1 include fibronectin and immunoglobulin regions, and using recombinant fragments to these regions we found that blocking the immunoglobulin domain at —30min, but not at +5.5hr, resulted in amnesia, while by contrast blocking the fibronectin domain at +5.5hr but not at —30min was amnestic (Scholey et al 1995). This biochemical version of a double dissociation experiment led us to postulate that it was the cell signalling function of L1, mediated via the immunoglobulin domain, which was engaged in the early phases of memory formation, whilst the fibronectin domains of NCAM and L1 were required in the deadherence/ readherence processes at the later time point. It is presumably at this time, 5-8hr downstream of the training event, whilst their epitopes on the external domains are open to attack, that antibody binding can occur and hence amnesia result.

Somewhat to our surprise, we have recently found that antibodies to two other transmembrane proteins also produce amnesia if injected 30 min pretraining, but not at later times. These are the lipid trafficking protein ApoE, and the cell-matrix adherent amyloid precursor protein, APP (Lancashire, Mileusnic and Rose, 1998; Mileusnic, Lancashire, Johnston and Rose, 1999). These effects may not be mutually independent, as there is evidence for interaction between APP and ApoE (Huber, Martin, Lofler and Moreau, 1993; Lancashire et al, 1998)and both may be involved in the transmembrane signalling processes that we have argued are required in the early stages of memory formation. As both proteins are implicated in Alzheimer’s disease, further elucidation of their mode of action in normal memory formation may provide clues as to the memory impairment characteristic of that condition.


Structural consequences


The longer-term consequence of this cascade is thus the modification of synaptic connectivity, detectable biochemically in terms of changes in the configuration and distribution of NCAM, amongst other synaptic markers. The presumed end point for memory storage is modulation of synaptic connectivity, by altering synaptic number, and/or relocating or structurally modifying existing synapses and dendritic spines. Such changes have been observed in a variety of paradigms, from enriched and impoverished environments to LTP and memory formation (Chang and Greenough, 1984: Bailey and Kandel, 1993). Might the biochemical changes we have observed in the chick IMHV following passive avoidance training also result in changes in synaptic morphology, detected by quantitative microscopy? In a series of studies, Stewart and his colleagues have been able to show changes in both pre- and post-synaptic elements. Thus 24hr post training there is increased dendritic spine density in projection neurons of the IMHV (Patel, Rose and Stewart, 1988), and at the same time changes in the numbers and dimensions of synaptic junctions, presynaptic boutons and synaptic vesicle number in both IMHV and lobus parolfactorius. These findings must however be contrasted with the observation by the Magdeburg group that, seven days following acoustic imprinting, there in a sharp decrease in spine number in the relevant forebrain region.(Wallhauser and Scheich, 1987). These results might be reconciled if we envisage that following training there is an initial efflorescence of spines, followed by a subsequent pruning, as argued, for instance, by Edelman (1985); but see Purves (1988). The increases in synapse number at 24hr found by Stewart are protein-synthesis dependent. However there are also transient changes in number detectable with an hour post training in the right IMHV, which may be attributable to synapse splitting rather than de novo synapse production (For reviews of these morphological changes see Stewart and Rusakov, 1995; Rose and Stewart, 1999).

  Factors affecting the salience of memory


Having pecked a bead coated in MeA, chicks avoid a similar but dry bead for at least 24hr subsequently. However, if the aversant is made less strong, by for instance using a 10% solution of MeA in alcohol, the birds peck and display a disgust reaction, but will avoid similar beads for only 6-9hr subsequently (Sandi and Rose 1994a; see also Burne and Rose, 1997). Although insofar as we have compared them, weak training initiates a similar set of synaptic transients to those produced in the strong version of the task, these are apparently not sufficient to result in gene expression, as glycoprotein synthesis does not occur. Our assumption is that the temporal relationship between the fading of the memory trace for the weak training beyond 6hr and the wave of glycoprotein synthesis that occurs at this time with strong training is not fortuitous (Rose, 1995b). However, there are a variety of factors which can affect the salience of this ‘weak learning’ experience, which result in memory being retained as for the strong learning.

Chicks are normally held in their pens in pairs, as this diminishes stress. If they are trained on 10% MeA, and then separated, stress levels increase, and retention persists for 24hr (Johnston and Rose, 1998). The normal training procedure is indeed stressful, as is shown by the fact that 5-10min following training chicks on the strong but not the weak version of the task there is an increase in plasma corticosterone levels (Sandi and Rose, 1997a). Further if corticosterone is injected into the IMHV just before or just after weak training, retention is also enhanced (Sandi and Rose, 1994a). The enhancing effects of stress can be blocked by injection of antagonists of glucocorticoid receptors into the IMHV, which is rich in such receptors. Blockade of these receptors is also amnestic for strong training (Sandi and Rose, 1994b), as is inhibition of peripheral corticosterone synthesis with metyrapone or aminoglutethimide (Loscertales, Rose and Sandi, 1997). Corticosterone has effects both at the membrane and genomically and it is relevant that we have shown that, independently of training, injections of corticosterone into the IMHV will enhance glycoprotein synthesis 6hr subsequently (Sandi, Rose, Mileusnic and Lancashire, 1995; Sandi and Rose, 1997b). As might be anticipated the effects of corticosterone are dose-dependent in the classic inverted-U form. 1g injected prior to weak training enhances retention, whilst higher doses do not, 1-5g injected prior to strong training diminishes retention.(Sandi and Rose, 1997a).

Neurotrophins also affect the salience of weak training. Recombinant brain-derived neurotrophic factor (BDNF), but not NGF or NT-3, injected just before or just after weak training, will enhance 24hr retention. Reciprocally, antibodies to BDNF are amnestic for strong training, amnesia setting in within 3hr (Johnston, Clements, and Rose, 1999). How BDNF may exert its effect is not clear. There are suggestions that it may act as a retrograde messenger, that it generates an intracellular signalling cascade via TrkB receptors, and that, like corticosterone, it may interact with GABAergic systems (see McKay, Purcell and Carew, 1999 for review). In this respect the fact that just as GABA agonists are amnestic for strong training, so antagonists enhance weak training (Clements and Bourne, 1996) may be significant.

These findings are of both theoretical and practical relevance. First, they remind us that although, especially under the influence of the neurophysiological observations of synaptic interactions during LTP, cellular theories of memory formation are heavily based on Hebbian models, memory is not just a pre/post synaptic event. Rather, whether any particular experience is learned or not depends on a much wider array of neural and peripheral factors, humoral and perhaps also immunological (see McGaugh, 1989, Damasio, 1994). The entire animal is thus involved in any learning experience. Second, together with the observations on ApoE and APP described in the previous section, they may point the way towards developing effective agents for therapeutic intervention in conditions of memory deficit.

  Memory beyond the IMHV


This review has focussed on the sequence of biochemical events occurring in the chick IMHV consequent on passive avoidance training, and I have argued that the cascade we have identified, leading as it does to measurable morphological changes in synaptic connectivity, is a necessary part of memory consolidation. Does this however mean that the IMHV contains some lasting representation of the association between bead and bitter taste, the elusive engram? A combination of electrophysiological and lesioning experiments which have been conducted in parallel with those described here, but are outside the remit of this review, makes clear that this is not simply the case (For review, Rose, 1999). Within the hours following training, biochemical changes occur in other brain regions than the left IMHV including the right IMHV and lobus parolfactorius, and the memory trace, if such it is, becomes both fragmented and redistributed. The IMHV seems to retain some aspects of the memory including a colour discrimination, whilst others, related perhaps to the size and shape of the bead, may be located to the lobus parolfactorius (Patterson and Rose 1992; Barber et al.1999). Again this points to the conclusion that learning and memory formation and retention engage not simply a discrete neuronal ensemble in a small brain region, but a much wider set of spatially and temporally dynamic processes, linked and given coherence by some form of binding mechanism (Freeman, 1999; Singer, 2000).

  Memory beyond the chick


I have argued that the chick may be at least one of God’s organisms for the study of the molecular processes of memory formation, but it is important to address the question of whether the processes we have observed are a special case, confined to either early learning, or to avian species. Whilst there are undoubtedly features which are specific to both, I would argue that our findings are of more general significance. The striking temporal parallels between the biochemical cascades found in the hippocampus following inhibitory avoidance training in rats, and in LTP, with those described here have been reviewed by Izquierdo and Medina (1997). The two ‘waves’ of glycoprotein synthesis following training have been found in a variety of adult mammalian species and tasks, including appetitive and inhibitory avoidance and motor learning in mice and rats (Grecksch and Matthies, 1980; Doyle et al 1992b). The antibodies originally prepared against chick glycoprotein species, notably NCAM, and which are amnestic when injected 5.5 hr post-training in our task, are also amnestic when injected at the same time in rats (Alexinsky, Przybyslawski, Mileusnic, Rose, and Sara, 1997; Roullet, Mileusnic, Rose, and Sara, 1997; and see also Doyle, et al. 1992b). However, in these cases amnesia does not set in until 48hr later, implying some more complex phase shifts in longer term memory. Indeed one of the clear implications of the cascade we have found in the chick is that the earlier heuristic model in which memory formation proceeds serially through a series of sequential short, intermediate and then finally robust long-term memory, presumably in the same small ensemble of neurons, is no longer sustainable, either for the chick or for other species (DeZazzo and Tully,1995). Furthermore the magnitude and diverse locations in space and time of the changes we have found following training on such a simple learning task demands that we reconceptualise our model of memory storage, moving from a fixed and linear view of memory formation to a more dynamic concept, involving large ensembles of cells differentially distributed in space and time (Rose, 1999).

It is clear from the diversity of receptor mediated events that can trigger LTP (eg NMDA and non-NMDA dependence) that we may expect to find that the processes that initiate synaptic transients are likely to be task and even species dependent. However, the intracellular and molecular housekeeping cascades that follow from these transients may well prove to be more universal, and the final involvement of CAMs in synaptic remodelling a common feature of many forms of memory. In elucidating these molecular mechanisms, and in its potential as a model system for the study of therapeutic approaches to memory impairment, the chick has earned its proper place amongst God’s organisms.



The research reviewed here has been supported by the UK Medical and Biotechonology and Biological Sciences Research Councils, the Wellcome Trust, the Royal Society and a collaborative programme funded by the European Science Foundation. I wish to thank my many colleagues in the Brain and Behaviour Research Group for the thirty years of collaboration reflected in this review and the references below.



Alexinsky, T, Przybyslawski, J, Mileusnic, R, Rose, SPR and Sara, SJ (1997) Antibody to day-old chick brain glycoprotein produces amnesia in adult rats Neurobiol. Learn. Mem. 67, 14-20

Ali, S.M., Bullock, S and Rose, S.P.R. (1988) Phosphorylation of synaptic proteins in chick forebrain: Changes with development and passive avoidance training. J. Neurochem. 50, 1579-1587

Andrew, RJ (ed) (1991) Neural and Behavioural Plasticity: the use of the Domestic Chick as a Model Oxford University Press

Andrew, RJ (1999) The differential role s of right and left sides of the brain in memory formation Behav. Brain Res 88, 289-296

Andrew, RJ and Rogers, LJ (1972) Testosterone, search behaviour and persistence Nature, 237, 343-346

Anokhin, K.V., Mileusnic, R, Shamakhina, I.Y. and Rose, S.P.R. (1991) Effects of early experience on c-fos gene expression in the chick forebrain. Brain Res. 544 101-107

Anokhin, K.V. and Rose, S.P.R. (1991) Learning-induced increase of immediate early gene messenger RNA in the chick forebrain. Eur. J. Neurosci. 3 162-167

Bailey, CH and Kandel, ER (1993) Structural changes accompanying memory storage Ann. Rev. Physiol. 55, 397-426

Barber, A., Gilbert, D.B. and Rose, S.P.R. (1989) Glycoprotein synthesis is necessary for memory of sickness-induced learning in chicks. Eur. J. Neurosci. 1, 673-677

Barber, TA, Howorth, PD, Klunk,AM and Cho, CC (1999) Lesions of the intermediate medial hyperstriatum ventrale impair sickness conditioned learning in day old chicks Neurobiol. Learn. Mem. 72, 128-141

Bateson, PPG (1966) The Characteristics and Context of Imprinting Biol. Rev. 41, 177-220

Bliss, TVP and Collingridge, GL (1993) A synaptic model of memory: Long-term potentiation in the hippocampus Nature, 361, 31-39

Bradley, PM, Burns,BD, King, TM and Webb, AC (1999) Electrophysiological correlates of past history: in vitro studies of the IMHV of domestic chick Bheav. Brain Res, 98, 261-266

Braun, K, Bock, J, Metzger, M, Jiang, S and Schnabel, R (1999) The dorsal neostriatum of the domestic chick: a structure serving higher associative functions Behav. Brain Res. 98, 211-218

Bullock, S., de Graan, P.N.E., Oestreicher, A.B., Gispen, W.H. and Rose, S.P.R. (1990) Identification of a 52kDa chick brain membrane protein showing changed phosphorylation after passive avoidance training as B50 (GAP 43). Neurosci. Res. Comm.6, 181-186

Burchuladze, R, and Rose, SPR (1992) Memory formation in day old chicks requires NMDA but not non-NMDA glutamate receptors. Eur.J.Neurosci. 4 533-538

Burchuladze, R, Potter, J and Rose, S.P.R. (1990) Memory formation in the chick depends on membrane-bound protein kinase C. Brain Res.535, 131-138

Burne, THJ and Rose, SPR (1997) Effects of training procedure on memory formation using a weak passive avoidance learning paradigm. Neurobiol. Learn. Mem. 68, 133-139

Chang, FLF and Greenough, WT (1984) Transient and enduring morphological correlates of synaptic activity and efficacy change in rat hippocampal slices Brain Res. 309, 35-46

Cherkin, A (1969) Kinetics of memory consolidation; role of amnestic treatment parameters. Proc Nat Acad Sci US 63, 1094-1101

Clements, MP and Bourne RC (1996) Passive avoidance learning in the day old chick is modulated by GABAergic agents Pharmacol. Biochem. Behav. 53, 629-634

Clements MP and Rose SPR (1996) Time dependent increase in release of arachidonic acid following passive avoidance training in the day-old chick J Neurochem. 67, 1317-1323

Clements, MP, Rose, SPR and Tiunova, A (1995).wConotoxin GVIA disrupts memory formation in the day old chick Neurobiol. Learn.Mem. 64 276-284

Csillag, A (1999) Striato-telencephalic and striato-tegmental circuits: relevance to learning in domestic chicks Behav. Brain Res. 98, 227-237

Crowe, SF, Zhao,W-Q, Sedman, GL and Ng, KT (1994) 2-deoxygalactose interferes with an intermediate processing stage of memory Behav. Neural Biol. 61, 206-213

Daisley, JN and Rose, SPR (1994) The effect of a passive avoidance task on the release of amino acids in vitro from the left IMHV of the day old chick. Biochem. Trans. 22, 160S

Daisley, JN, Gruss, M, Rose, SPR and Braun, K (1998) Passive avoidance training and recall are associated with increased glutamate levels in the IMHV of the day old chick Neural Plasticity, 6,41-52

Damasio, AR (1994) Descartes Error: emotion, reason and the human brain.London, Macmillan

Davies, HP and Squire,LR (1984) Protein synthesis and memory: a review Psychol. Bull. 96, 518-559

DeZazzo, J and Tully, T (1995) Dissection of memory formation. From behavioral pharmacology to molecular genetics. Trends Neurosci. 18, 212-217

Doyle, E, Nolan, P, Bell, R and Regan, CM (1992a) Neurodevelopmental events underlying information acquisition and storage Network, 3, 89-94

Doyle, E, Nolan, P, Bell, R and Regan, CM (1992b) Intraventricular infusions of anti-neural cell adhesion molecule in a discrete post-training period impair consolidation of a passive avoidance response in the rat J, Neurochem. 59, 1570-1573

Edelman, GM (1985) Cell adhesion and the molecular process of morphogenesis Ann. Rev Biochem. 54, 135-169

Freeman, FM and Rose, SPR (1995) Two time windows of anisomycin-induced amnesia for passive avoidance training in the day old chick Neurobiol. Learn. Mem .63, 291-295

Freeman, W (1999) How Brains Make up their Minds,Weidenfeld and Nicolson, London

Gibbs, ME and Ng, KT (1977) Psychobiology of memory: Towards a model of memory formation. Behav, Rev. 1, 113-136

Gibbs, ME, Ng, KT and Crowe, S (1991) Hormones and the timing of phases of memory formation. In Andrew, RJ (ed) Neural and Behavioural Plasticity: the use of the Domestic Chick as a Model Oxford University Press, pp440-455

Gilbert, D.B., Patterson, T.A., and Rose, S.P.R. (1989) Midazolam induces amnesia in a simple one-trial maze-learning task. Pharmacol. Biochem. Behav. 34, 439-442

Goelet, P, Castelluci, S, Schachner, S and Kandel, ER (1986) The long and the short of long-term memory — a molecular framework Nature 322, 419-423

Grecksch, G and Matthies, H-J (1980) Two sensitive periods for the amnestic effect of anisomycin Pharmacol. Biochem Behav. 12, 663-665

Gruss, M and Braun, K (1996) Stimulus-evoked increase of glutamate in the mediorostral neostratum/hyperstriatum ventrale of domestic chick after auditory imprinting: an in vivo microdialysis study. J. Neurochem. 66, 1167-1173

Holscher, C and Rose, SPR (1993) Inhibiting synthesis of the putative retrograde messenger nitric oxide results in amnesia in a passive avoidance task in the chick. Brain Res. 619,189-194

Holscher, C and Rose, SPR (1994) Inhibitors of phospholipase A2 produce amnesia for a passive avoidance task in the chick. Behav.Neural Biol. 61, 225-232

Horn, G (1985) Memory, Imprinting and the Brain , Oxford University Press

Huber, G, Martin, JR, Lofler, J and Moreau, J-L (1993) Involvement of amyloid precursor protein in memory formation in the rat:L an indirect antibody approach Brain Res. 603, 348-352

Izquierdo, I. & Medina, J.H.(1997) Memory formation: the sequence of biochemical events in the hippocampus and its connection to activity in other brain structuresNeurobiol. Learn. Memory 68, 285-316

Johnston ANB, Burne, THJ and Rose, SPR (1999) Observation learning in day-0old chicks using a one-trial passive avoidance learning paradigm. Anim. Behav. 56, 1347-1353

Johnston, ANB and Rogers, LJ (1998) Right hemisphere involvement in imprinting memory revealed by glutamate treatment. Pharmacol. Biochem. Behav. 60, 863-871

Johnston, ANB and Rose, SPR (1998) Isolation-stress induced facilitation of passive avoidance memory in the day-old chick Behav. Neurosci. 112, 1-8

Johnston, ANB, Clements, MP and Rose, SPR (1999) Role of brain-derived neurotrophic factor and presynaptic proteins in passive avoidance learning in day old domestic chicks. Neurosci. 66, 1033-1042

Konishi,M (1989) Bird song for neurobiologists Neuron, 3, 541-549

Kossut, M and Rose, SPR (1984) Differential 2-deoxyglucose uptake into chick brain structures during passive avoidance training Neurosci. 12, 971-977

Krebs, JR, Sherry, DF, Healy, SD, Perry, VH and Vaccarino, AL (1989) Hippocampal specialisation of food storing birds Proc. Natt. Acad. Sci. US, 86, 1388-1392

Lancashire, C, Mileusnic, R and Rose, SPR (1998) Apolipoprotein antibodies affect the retention of passive avoidance memory. Neural Plasticity, 6, 29-40

Lorenz, K (1935) Der Kumpan in der Umwelt des Vogels J.Ornithol.. 83, 137-213; 289-413

Loscertales, M. Rose, SPR and Sandi, C (1997) The corticosteroid synthesis inhibitors metyrapone and aminoglutethimide block long-term memory for a passive avoidance task in day-old chicks Brain Res., 769, 357-361

Lossner, B. and Rose, S.P.R.(1983) Passive avoidance training increases fucokinase activity in right forebrain base of day-old chicks. J. Neurochem. 41 , 1357-1363.

Luthi, A, Laurent, JP, Figurov, A, Muller,D and Schachner, M (1994) Hippocampal long term potentiation and neural cell adhesion molecules L1 and NCAM Nature, 372, 777-779

McCabe, BJ and Nicol, AU (1999) The recognition memory of imprinting: biochemistry and electrophysiology,Behav. Brain Res.98, 253-260

McKay, SE, Purcell, AL and Carew, TJ (1999) Regulation of synaptic function by neurotrophic factors in vertebrates and invertebrates: implications for development and learning. Learn. Mem. 6, 193-215

McGaugh, JL (1989) Involvement of hormonal and neuromodulatory systems in the regulation of memory storage. Ann. Rev. Neurosci. 12, 255-287

Marler, P (1991) Song learning behavior: the interface with neuroethology Trends Neurosci. 14, 199-206

Matthies, H-J (1989) In search of the cellular mechanisms of memory Prog. Neurobiol.32, 277-349

Mileusnic, R, Anokhin, K and Rose, SPR (1996) Antisense oligodeoxynucleotides to c-fos are amnestic for passive avoidance in the chick NeuroReport, 7, 1269-1272

Mileusnic, R, Lancashire, C and Rose, SPR (1999) Sequence specific impairment of memory formation by NCAM antisense oligonucleotides Learning and Memory 6,120-127

Mileusnic, R, Lancshire, C, Johnston, ANB and Rose, SPR (1999) Antibodies and antisense to the amyloid precursor protein (b-APP) prevent memory consolidation:RERMS peptide rescues in preparation

Mileusnic, R, Rose, SPR, Lancashire, C and Bullock, S (1995) Characterisation of antibodies specific for chick brain NCAM which cause amnesia for a passive avoidance task J. Neurochem.64 2598-2605

Ng, KT and Gibbs, ME (1991) Stages in memory formation: a review. In Andrew, RJ (ed) Neural and Behavioural Plasticity: the use of the Domestic Chick as a Model Oxford University Press, pp351-369

Nottebohm, F (1970) Ontogeny of bird song Science, 167, 950-956

Nottebohm, F (1980) Brain pathways for vocal learning in birds: A review of the first ten years Progress in Psychobiology and Physiological Psychology, 9, 85-124

Patel, S.N., Rose, S.P.R., and Stewart, M.G. (1988) Training induced dendritic spine density changes are specifically related to memory formation processes in the chick, gallus domesticus. Brain Research 463, 168-173

Patterson, TA, Alvarado, MC, Rosenzweig, MR and Bennett, EL (1988) Time courses of amnesia development in two different areas of the chick brain Neurochem. Res. 10, 1071-1081

Patterson, TA and Rose, SPR (1992) Memory in the chick: multiple cues, distinct brain locations. Behavioural Neuroscience, 106, 465-470

Pollerberg, GE, Sadoul, R, Goridis, C and Schachner, M (1985) Selective expression of the 180-kD component of the neural cell adhesion molecule NCAM during development J. Cell Biol. 101. 1921-1929

Popov, NS, Schulzeck, S, Pohle, W and Matthies, H-J (1981) Changes in the incorporation of 3H-fucose into rat hippocampus after acquisition of a brightness discrimination reaction: an electrophoretic study Neuorscience, 5, 161-165

Purves, D (1988) Body and Brain: a trophic theory of neural connections Harvard University Press, Cambridge Ma

Regolin, L and Rose, SPR (1999) Long-term memory for a spatial task in young chicks Anim. Behav. 57, 1185-1191

Rickard, NS, Poot, AC, Gibbs, ME and Ng, KT (1994) Antagonism of iether NMDA or non-NMDA glutamate receptors impairs long-term memory consolidation in the day-old chick Behav. Neural Biol. 62, 33-40

Rickard, NS.; Ng, KT. & Gibbs, ME. (1998). Further support for nitric oxide-dependent memory processing in the day-old chick. Neurobiol. Learn. Memory. 69; 79-86.

Rogers, LJ (1995) The Development of Brain and Behaviour in the Chicken CAB International, Wallingford, UK

Rogers, LJ and Deng, C (1999) Light experience and the lateralisation of two visual pathways in the chick. Behav. Brain Res. 88 277-288

Rose, SPR (1995a) Glycoproteins and memory formation Behav. Brain Res. 66 73-78

Rose, SPR (1995b) Cell adhesion molecules, glucocorticoids and memory Trends Neurosci. 18, 502-506

Rose, SPR (1999) Time dependent processes in memory revisited Behav. Neurosci. In press

Rose, S.P.R. and Harding, S.(1984) Training increases 3H Fucose incorporation in chick brain only if followed by memory storage. Neurosci 12, 663-667.

Rose, S.P.R. and Jork, R.(1987) Long-term memory formation in chick is blocked by 2-deoxygalactose, a fucose analogue. Behav.Neural.Biol. 48, 246-258

Rose, SPR and Stewart, MG (1999) Cellular correlates of memory formation in the chick following passive avoidance training Behav. Brain Res. 98, 237-243

Roullet. P. Mileusnic, R, Rose, SPR and Sara, SJ (1997) Neural cell adhesion molecules play a role in rat memory formation in appetitive as well as aversive tasks Neuroreport, 8, 1907-1911

Rusakov, DA, Davies, HA, Krivko, IM, Stewart MG and Schachner, M (1994) Training alters patterns of PSA-NCAM in chick forebrain membranes NeuroReport 5, 2467-2473

Salinska, EJ, Bourne,RC and Rose, SPR (2000) Long-term memory formation in the chick requires mobilisation of ryanodine sensitive intracellular calcium stores. Neurobiology of Learning In Press

Salinska, EJ, Chaudhury, D, Bourne, RC and Rose, SPR (1999) Passive avoidance training results in increased responsiveness of synaptosomal voltage and ligand gated channels in chick brain Neuroscience In press

Sandi, C, and Rose, S.P.R. (1994a) Corticosterone enhances long-term retention in one day old chicks trained in a weak passive avoidance learning paradigm. Brain Res.647, 106-112

Sandi, C and Rose, SPR (1994b) Corticosteroid receptor antagonists are amnestic for passive avoidance learning in day-old chicks Eur. J. Neurosci. 6, 1292-1297

Sandi, C and Rose, SPR (1997a) Training-dependent biphasic effects of corticosterone in memory formation for a passive avoidance tasks in chicks. Psychopharmacology 133, 152-160

Sandi, S and Rose, SPR (1997b) Protein synthesis and fucosylation-dependent mechanisms in corticosterone facilitation of long-term memory in the chick Behav. Neurosci.111, 1-7

Sandi, C, Rose, SPR, Mileusnic, R and Lancashire, C (1995) Corticosterone facilitates long-term memory formation via enhanced glycoprotein synthesis Neuroscience, 69, 1087-1093

Scheich, H (1987) Neural correlates of auditory filial imprinting J. Comp. Physiol. 161, 605-9

Scholey, AB, Mileusnic, R, Schachner, M and Rose, SPR (1995) A role for a chicken homolog of the neural cell adhesion molecule L1 in consolidation of memory for a passive avoidance task Learn. Mem. 2,17-25

Scholey, AB, Rose, SPR, Zamani, MR, Bock, E and Schachner, M (1993) A role for the neural cell adhesion molecule in a late consolidating phase of glycoprotein synthesis 6hr following passive avoidance training of the young chick Neuroscience 55, 499-509

Serrano, PE,Beniston, DS, Oxonian, MG, Rodriguez, WA, Rosenzweig,MR and Eennett, EL (1994) Differential effects of protein kinase inhibitors and activators on memory formation in the 2-day old chick Behav. Neural Biol. 61, 60-72

Singer, W (2000) In Changing Concepts of Nature at the Millennium, Cambridge University Press, forthcoming

Skibo, GG, Davies, HA, Rusakov, DA, Stewart, MG and Schachner, M (1993) Increased immunogold labelling of NCAM isoforms in synaptic active zones of the chick striatum 5-6hr after one-trial passive avoidance training Neuroscience, 82, 1-5

Solomonia, RO, McCabe, BJ and Horn, G (1995) A time-dependent, larning specific increase in the amount of neural cel adheison molecules (NCAMs) in chick forebrain after visual imprinting J.Physiol., 485, P8

Steele, RJ and Stewart, MG (1995) Involvement of AMPA receptors in maintenance of memory for a passive avoidance task in day-old domestic chicks (Gallus domesticus) Eur. J. Neurosci.,7, 1297-1304

Steele, RJ, Stewart, MG and Rose, SPR (1995) Increases in NMDA receptor binding are specifically related to memory formation for a passive avoidance task in the chick: a quantitative autoradiographic study. Brain Res. 674, 352-356

Stewart , MG, Bourne, RC and Steele, RJ (1992) Quantitative autoradiographic demonstration of changes in binding to NMDA sensitive 3H glutamate and 3H MK801 but not 3H AMPA receptors in the chick forebrain 30 minutes after passive avoidance training Eur. J. Neurosci. 4, 936-943

Stewart, MG and Rusakov, DA (1995) Morphological changes associated wirth stages of memory formation in the chick following passive avoidance training Behav. Brain Res. 12, 21-28

Tiunova, A, Anokhin, KV and Rose, SPR (1998) Two critical periods of protein and glycoprotein synthesis in memory consolidation for visual categorisation learning in chicks. Learn. Mem. 4, 401-410

Wallhauser, E and Scheich, H (1987) Auditory imprinting leads to differential 2-deoxyglucose uptake and dendritic spine loss in the chick rostral forebrain Devel. Brain Res. 31, 29-44