Experience-dependent changes in the hippocampusof domestic chicks: a model for spatial memory

Rafael Freire1 and Heng-Wei Cheng2

1Centre for Neuroscience and Animal Behaviour, School of Biological Biomedical and Molecular Sciences, University of NewEngland, Armidale, NSW 2351, Australia2United States Department of Agriculture – Animal Research Services, Poultry Building, West Lafayette, IN47906, USA

Keywords: behaviour, chicken, development, memory, plasticity

Abstract

In the domestic chicken, providing visual barriers for a brief period early in life has been found to improve spatial memory [R. Freireet al. (2004) Animal Behaviour, 67, 141–150]. In the present study we compared the structure of neurons in the hippocampus andneostriatum in chicks reared with or without visual barriers. From 8 to 16 days of age, chicks were reared in pens either with twowooden screens (Treatment E) or with no screens (Treatment C). At 16 days of age, chicks were anaesthetized, perfusedintracardially and brain samples collected and stained using a Golgi–Cox technique. Morphometric analysis revealed that themultipolar projection neurons of Treatment E chicks had longer dendrites (anova, F1,14 ¼ 7.4, P < 0.05) and had more spines per20 lm of dendrite (SLD; anova, F1,14 ¼ 10.6, P < 0.01) than those of Treatment C chicks. By contrast, no evidence was found thatrearing treatment differentially influences dendrite length or SLD in the neostriatum, suggesting that the above environment-inducedchanges may be specific to the hippocampus. Multipolar projection neuron dendrites of the right hemisphere were longer (anova,F1,14 ¼ 36.4, P < 0.0001) and had more spines (anova, F1,14 ¼ 8.8, P < 0.05) than dendrites of the left hemisphere, supportingprevious findings that the right hemisphere of chickens is predominantly involved in spatial processing. We conclude that the chickenprovides a useful model for the study of developmental plasticity in brain and behaviour, partly because the possibility of rearingchicks in isolation and imprinting them on an artificial object provides a means of accurately manipulating early experience.

Introduction

As in mammals, the avian hippocampus appears to be centrallyinvolved in the processing and storage of spatial information(Colombo & Broadbent, 2000). For example, lesioning of thehippocampus in passerine birds impairs spatial memory but has littleeffect on other cognitive tasks (Hampton & Shettleworth, 1996;Fremouw et al., 1997), and spatial memory in hippocampus-lesionedzebra finches can be reversed by transplantation of embryonic tissueinto the hippocampus (Patel et al., 1997). Recently, Tommasi et al.(2003) found that lesion of the right hippocampus of domesticchickens impaired search based on relative positional cues, whereaslesion of the left hippocampus had little effect on spatial memory.

Scientific evidence mainly involving rodent models has largelysupported the theory that experience plays a crucial role in thedevelopment of brain and behaviour (for review see Rosenzweig &Bennett, 1996). Such evidence has far-reaching implications, becausefor example the development of abnormal behaviour and brainfunction in laboratory rodents may provide misleading scientificfindings in medical research (Wurbel, 2001). One limitation of therodent model for research on experience-induced modulation in brainand behaviour is that the rodents’ dependency on maternal bondsmakes it difficult to manipulate and replicate accurately earlyexperience. Indeed, it now appears that maternal behaviour shapesbrain and behaviour and that this process is important for the

subsequent survival of the young rat (Francis & Meaney, 1999;Francis et al., 1999).Domestic chickens can be hatched and reared in isolation and

imprinted on an artificial object that provides a useful means ofcontrolling and manipulating important environmental factors.Domestic chicks in semi-natural conditions actively move out ofsight of the mother on day 11 (Vallortigara et al., 1997), a periodwhich also corresponds with a shift to right hemisphere dominance(which shows an advantage in spatial processing, see above; Rogers,1995). Recently, Freire et al. (2004) found that chicks reared inenvironments with visual barriers added at around 11 days of age hadbetter spatial memory, as determined by visual displacement anddetour tests, than chicks reared without visual barriers. Clearly,determining whether the above experience-induced changes in spatialmemory are related to similar changes in the structure of thehippocampus in the chicken would present an alternative (avian)model in the study of experience-induced changes in brain andbehaviour.The aim of the present study was to test the hypothesis that

experience of visual barriers from 8 to 16 days of age leads to greaterdevelopment in the hippocampus relative to chicks provided with novisual barriers (controls). Golgi staining was used to evaluatemultipolar projection neuron structure in the hippocampus, becausethese cells are widespread and easily identified in the hippocampus,and has previously proved an effective method for revealing neuronsand fibres in the chick hippocampus (Tombol et al., 2000). Thegenerality of experience-induced changes in brain structure wasexamined by evaluating a control area, the neostriatum, which is

Correspondence: Dr R. Freire, as above.E-mail: rfreire2@une.edu.au

Received 26 April 2004, revised 27 May 2004, accepted 10 June 2004

European Journal of Neuroscience, Vol. 20, pp. 1065–1068, 2004 ª Federation of European Neuroscience Societies

doi:10.1111/j.1460-9568.2004.03545.x

important in imprinting but apparently not in spatial processing (Bock& Braun, 1999).

Materials and methods

The subjects were 18 broiler chicks (unsexed chicks, Pine ManorHatchery, IN, USA) obtained as fertile eggs at 18 days of incubation.Chicks were reared in isolation from about 2 h after hatching to 7 daysof age in a cardboard box measuring 30 · 25 cm and 30 cm high, at atemperature of 35 �C and a light–dark cycle of 12 : 12 h. In order toencourage pecking and eating, the floor of the box was lined withwhite paper and sprinkled with chick starter crumbs that wereperiodically tapped with a round dowel. Additional starter crumbs andwater were available ad libitum from clear Perspex Petri dishes. Ayellow tennis ball was suspended 10 cm above the floor by a string inthe centre of the box to provide an imprinting stimulus. The aboveconditions exactly replicated those of Freire et al. (2004).At 8 days of age, chicks were paired with another identically reared

chick of the same age (social companion) and marked with blue spraypaint on the back and randomly assigned to two rearing treatments.Each treatment consisted of wood shavings in a box measuring55 · 40 cm and 60 cm high. A drinker was placed in one corner sothat birds could not walk around it, and starter crumbs placed in ashallow dish to provide ad libitum food and water. Temperature wasmaintained at 30 �C and the light–dark cycle was 12 : 12 h. A yellowtennis ball was suspended in the centre of the box. Nine chicks werereared in boxes with two 20 · 20-cm wooden screens placed centrallyand parallel to the shorter side of the box, 10 cm from the ball suchthat the ball was between the two screens (Treatment E; Fig. 1a). Theremaining nine chicks were reared in boxes without screens (Treat-ment C; Fig. 1b).

At 16 days of age, subjects were removed from the rearing boxes anddeeply anaesthetized with sodium pentobarbitol (30 mg ⁄ kg) andperfused intracardially with 1000 mL 0.9# sodium chloride. Brainsamples were collected and placed in Golgi–Cox solution for 14 days(according to the procedure ofVan der Loos, 1956). Sampleswere coded(with a random number from 1 to 18) so that the gathering of data wasperformed blind. After cell staining, the tissues were dehydrated, the leftside marked with a cut and serially sectioned at 200 lm in the frontalplane. Tissue sections were serially mounted onto slides and showedlarge numbers of darkly impregnated striatal cell bodies and dendritesagainst a relatively clear background. All animal procedures wereapproved by Purdue University’s Animal Ethics Committee.

Morphometric analysis

The hippocampus and neostriatum were identified under low magni-fication (·20) using an Olympus BX60 microscope. Sections of the

hippocampus approximately 400 lm anterior to the anterior commis-sure and of the neostriatum showing the anterior tractus septomesen-cephalicus were selected for further examination (Kuenzel & Masson,1988). Thirty multipolar projection neurons from the hippocampus(we did not attempt to differentiate between the layers of thehippocampus; Tombol et al., 2000) and 30 neurons from theneostriatum of each hemisphere were analysed under higher magni-fication. An intact and fully impregnated dendrite from a predeter-mined quadrant of each neuron was selected if it was unobscured byneighbouring striatal cells and could be traced to the tip (characterizedby tapering and the absence of spines). A lucent drawing tube wasused to superimpose an image onto paper for tracing the dendrite andall its spines at a magnification of ·1640. The dendrite length wasmeasured and transformed to actual length.

Statistical analysis

Mean number of spines was calculated for each 20-lm segment ofdendrites from a minimum of five neurons. Mean dendrite length andspine linear density (SLD) were analysed by an anova using a split plotdesign, with rearing treatment as a between-subjects comparison andhemisphere as a within-subject comparison (SPSS Base 10.0, SPSS inc,Chicago, IL,USA).Dendritemeasurements from the hippocampuswerenot obtained from two chicks (one from each treatment) as thehippocampus, which lies on the periphery and near the midline, becameseparated during sectioning and was not recovered.

Results

Dendrite length

Dendrites of the multipolar projection neurons of the hippocampuswere longer in Treatment E chicks than in Treatment C chicks(anova, F1,14 ¼ 7.4, P < 0.05; Fig. 2). Hippocampus dendrites of theright hemisphere were longer than dendrites of the left hemisphere(84.9 ± 3.0 lm and 74.9 ± 3.0 lm, respectively; anova,F1,14 ¼ 36.4, P < 0.0001). By contrast, no evidence was found thatthe length of the neostriatum dendrites differed between the two

a) Treatment E b) Treatment C

Fig. 1. The apparatus used to house (a) Treatment E and (b) Treatment Cchicks from 8 days of age. The figures show the imprinting stimulus suspendedin the middle of the rectangular pens. Treatment E contained two woodenscreens placed either side of the imprinting stimulus as shown.

Fig. 2. Mean (± SEM) length of the dendrites of the left (grey bars) and right(open bars) )hemisphere of the hippocampus and neostriatum. Dendrites ofmultipolar projection neurons of the hippocampus were longer in Treatment Echicks than in Treatment C chicks (anova, F1,14 ¼ 7.4, P < 0.05).

1066 R. Freire and H.-W. Cheng

ª 2004 Federation of European Neuroscience Societies, European Journal of Neuroscience, 20, 1065–1068

rearing treatments (anova, F1,16 ¼ 0.07, NS) or between the left andright hemispheres (anova, F1,16 ¼ 1.5, NS). No significant interac-tions were found between treatment and hemisphere for the hippo-campus (anova, F1,14 ¼ 0.9, NS) or neostriatum (anova,F1,16 ¼ 0.8, NS).

Spine linear density

There were significantly more spines per 20 lm of dendrite ofmultipolar projection neurons in Treatment E chicks than in TreatmentC chicks (Fig. 3; anova, F1,14 ¼ 10.6, P < 0.01). Figure 3 suggeststhat the number of spines per 20 lm of dendrite was greater for all 20-lm segments except the segment nearest to the cell body. SLD wasalso greater in the right hemisphere of the hippocampus than in the left(Fig. 3; anova, F1,14 ¼ 8.8, P < 0.05). By contrast, no evidence wasfound that SLD for neostriatum dendrites was influenced by treatment(Fig. 4; anova, F1,16 ¼ 0.3, NS) or differed between hemispheres(anova, F1,16 ¼ 0.3, NS).

No evidence was found of increased changes in either thehippocampus or the neostriatum due to the rearing treatment in theright hemisphere relative to the left in either dendrite length (seeabove) or SLD (treatment–hemisphere interaction, hippocampus,anova, F1,14 ¼ 1.7, NS; neostriatum, anova, F1,16 ¼ 3.1, NS).

Discussion

We found that the multipolar projection neurons of the hippocampusof chicks reared with visual barriers had longer dendrites with morespines than those of chicks reared without visual barriers. Onepossibility is that the different rearing conditions have an effectunrelated to spatial memory, with perhaps the most likely possibilitythat the rearing treatments, as they influenced the degree of visualexposure to the imprinting stimulus, may have provided differentdegrees of imprinting. It is unlikely, however, that the above findingsare related to a differential degree of imprinting in the twoexperimental groups for several reasons. First, neural changes inresponse to imprinting are maximal in the first 7 days of life andrelatively minor thereafter (Rogers, 1995). Second, we found nodifference in spine density in the neostriatum between rearingtreatments even though imprinting has previously been shown toreduce spine density in the neostriatum (Bock & Braun, 1999). Third,although the intermediate medial hyperstriatum ventrale (IMHV) isinvolved in imprinting and has connections to the hippocampus, theabsence of a correlation between neuronal action in the IMHV andhippocampus suggests that the hippocampus has no role in imprinting(Nicol et al., 1998). In fact, Nicol et al. (1998) found that the responseof many hippocampal neurons was dependent on the distance to thestimulus rather than the type of stimulus, indicating that thehippocampus plays a role in the visual estimation of distance. Thusin view of the known involvement of the avian hippocampus in spatialprocessing (see Introduction) and the recent findings that experience ofvisual barriers improves spatial memory (Freire et al., 2004), it seemsprobable that the experience-induced changes in the hippocampusreported here were related to spatial memory. Additionally, theabsence of experience-induced changes in the neostriatum strengthensthe assertion that the neural changes observed were confined to thehippocampus.The present findings that the dendrites of the right hippocampus

were longer and had more spines than those of the left hippocam-pus support Rashid & Andrew’s (1989) assertion that the righthemisphere is predominantly involved in spatial processing in thechicken. More recently, Tommasi & Vallortigara (2001) found thatthe right hemisphere was able to encode information on the relativeposition of objects whereas the left hemisphere was not. Usinghippocampus-lesioned chicks in a similar test to the latterexperiment, Tommasi et al. (2003) found that an intact right

0

2

4

6

8

10

12

0-20 20-40 40-60 60-80 80-100

Distance from cell body (microns)

Num

ber

ofsp

ines

Treatment E, Right Treatment E, Left Treatment C, Right Treatment C, Left

Fig. 3. Mean (± SEM) number of spines per 20 lm of hippocampus dendrite in the left and right hemisphere of Treatment E and Treatment C chicks.

0

2

4

6

8

10

12

14

0-20 20-40 40-60 60-80 80-100 100-120

Distance from cell body (microns)

Num

ber

ofsp

ines

Treatment E Treatment C

Fig. 4. Mean (± SEM) number of spines per 20 lm of neostriatum dendrite inTreatment E and Treatment C chicks. Mean of the left and right hemispheresare combined due to the similarity between the lines.

Chicken spatial memory and hippocampus 1067

ª 2004 Federation of European Neuroscience Societies, European Journal of Neuroscience, 20, 1065–1068

hippocampus was necessary for encoding relative positional infor-mation. It therefore seems that in the chicken, as in mammals andhumans, the right hippocampus is structurally distinct and hasfunctional advantages over the left hippocampus in spatial process-ing.In semi-natural conditions, domestic chicks show a peak in

moving out of sight of the mother at 11 days of age (Workman &Andrew, 1989), an action that would provide the chicken withexperience of the mother being out of sight (occluded). The additionof wooden barriers may, however, have different effects other thanproviding experience of occlusion, such as providing experience ofnavigating around a barrier. It is therefore impossible at this stage todetermine unequivocally whether experience of occlusion per seleads to the observed changes in the hippocampus. It may be that theaddition of wooden barriers provided other, perhaps more general,experiences that lead to the observed changes in the hippocampus.Nonetheless, it should be possible to control for alternativeexplanations, such as by using clear screens (as in Freire et al.,2004) and in this respect we propose that the chicken offers amethodologically useful model in understanding the role of criticalstimuli in shaping spatial memory.We found no support for our prediction that the addition of

visual barriers would lead to greater effects in the right comparedwith the left hippocampus (as would be indicated by a significantrearing treatment and hemispheric side interaction). In the domesticchicken there is a shift in hemispheric dominance towards the rightside on day 11 (Rogers & Ehrlich, 1983). At this stage we do notknow the extent to which the change observed in the hippocampusin our study was confined to day 11 (i.e. how specific this sensitivephase might be). However, the absence of a significant interactionbetween rearing treatment and hemisphere in the present studysuggests that this experience is not necessary in the development ofbrain laterality. Thus although spatial memory may predominantlyinvolve the right hemisphere, experience important in the develop-ment of spatial memory appears to influence development of bothsides of the brain. It should, however, be stressed that there arelimitations to the present study and that other variables not exploredhere may show brain asymmetry. Clearly, further studies arenecessary to determine the extent to which experience-inducedchanges in spatial memory are restricted to the right hemisphereand to day 11 of age.In conclusion, experience-induced changes in brain morphology

reported here suggest that early experience leads to changes in thehippocampus that appear to be related to the development of spatialmemory. Our failure to find similar environment-induced changes inthe neostriatum raises the possibility that the observed brain changeswere reasonably specific to the hippocampus. Differences between theleft and right hippocampus reported here are consistent with previousfindings suggesting that the right hemisphere in the chick ispredominantly involved in spatial processing. We propose that thedomestic chick, with its apparent sensitive phases for development,hemispheric specialization and behavioural and neural plasticity indevelopment, provides a useful model for investigating experience-induced changes in brain and behaviour.

Acknowledgements

We thank Pete Singleton for dissecting the chick brains, Andrea Quigley fordrawing and measuring neurons and Christine Nicol, Sue Healy and LesleyRogers for helpful contributions to this study. We also thank Purdue University,the USDA-ARS and the University of New England for financial support.

References

Bock, J. & Braun, K. (1999) Filial imprinting in domestic chicks is associatedwith spine pruning in the associative area, dorsocaudal neostriatum. Eur. J.Neurosci., 11, 2566–2570.

Colombo, M. & Broadbent, N. (2000) Is the avian hippocampus a functionalhomologue of the mammalian hippocampus. Neurosci. Biobehav. Rev., 24,465–484.

Francis, D., Diorio, J., Liu, D. & Meaney, M.J. (1999) Nongenomictransmission across generations of maternal behaviour and stress responsesin the rat. Science, 286, 1155–1158.

Francis, D. & Meaney, M.J. (1999) Maternal care and the development of stressresponses. Curr. Opin. Neurobiol., 9, 128–134.

Freire, R., Cheng, H.-W. & Nicol, C.J. (2004) Development of spatial memoryin occlusion-experienced domestic chicks. Anim. Behav., 67, 141–150.

Fremouw, T., Jackson Smith, P. & Kesner, R.P. (1997) Impaired place learningand unimpaired cue learning in hippocampal-lesioned pigeons. Behav.Neurosci., 111, 963–975.

Hampton, R.R. & Shettleworth, S.J. (1996) Hippocampal lesions impairmemory for location but not color in passerine birds. Behav. Neurosci., 110,831–835.

Kuenzel, W.J. & Masson, M. (1988) A Sterotaxic Atlas of the Brain of theChick. The Johns Hopkins University Press, Baltimore, MD.

Nicol, A.U., Brown,M.W.&Horn, G. (1998)Hippocampal neuronal activity andimprinting in the behaving domestic chick. Eur. J. Neurosci., 10, 2738–2741.

Patel, S.N., Clayton, N.S. & Krebs, J.R. (1997) Hippocampal tissue transplantsreverse lesion-induced spatial memory deficits in zebra finches (Taeniopygiaguttata). J. Neurosci., 17, 3861–3869.

Rashid, N. &Andrew, R.J. (1989) Right hemisphere advantage for topographicalorientation in the domestic chick. Neuropsychologia, 27, 937–948.

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

Rogers, L.J. & Ehrlich, D. (1983) Asymmetry in the chicken forebrain duringdevelopment and a possible involvement in supraoptic decussation.Neurosci. Lett., 37, 123–127.

Rosenzweig,M.R. & Bennett, E.L. (1996) Psychobiology of plasticity: effects oftraining and experience on brain and behavior. Behav. Brain Res., 78, 57–65.

Tombol, T., Davies, D.C., Nemeth, A., Sebesteny, T. & Alpar, A. (2000) Acomparative study of chicken (Gallus domesticus) and homing pigeon(Columba livia) hippocampus. Anat. Embryol., 201, 85–101.

Tommasi, L., Gagliardo, A., Andrew, R.J. & Vallortigara, G. (2003) Separateprocessing mechanisms for encoding of geometric and landmark informationin the avian hippocampus. Eur. J. Neurosci., 17, 1895–2702.

Tommasi, L. & Vallortigara, G. (2001) Encoding of geometric and landmarkinformation in the left and right hemispheres of the avian brain. Behav.Neurosci., 115, 602–613.

Vallortigara, G., Andrew, R.J., Sertoli, L. & Regolin, L. (1997) Sharply timedbehavioural changes during the first 5 weeks of life in the domestic chicken(Gallus gallus). Bird Behav., 12, 29–40.

Van der Loos, H. (1956) Une combinaison de deax Vieilles methodshistologiques pour le system nerveux ventral. Mschr. Psychiatr.–Neurol.,132, 330–334.

Workman, L. & Andrew, R.J. (1989) Simultaneous changes in behaviour and inlateralization during the development of male and female domestic chicks.Anim. Behav., 38, 596–605.

Wurbel, H. (2001) Ideal homes? Housing effects on rodent brain and behaviour.J. Neurosci., 24, 207–211.

1068 R. Freire and H.-W. Cheng

ª 2004 Federation of European Neuroscience Societies, European Journal of Neuroscience, 20, 1065–1068