Doi:10.1016/j.neurobiolaging.2004.02.014

Learning ability in aged beagle dogs is preserved by behavioral enrichment and dietary fortification: a two-year longitudinal study N.W. Milgram , E. Head , S.C. Zicker , C.J. Ikeda-Douglas , H. Murphey , B. Muggenburg , C. Siwak , D. Tapp , C.W. Cotman a University of Toronto, Division of Life Sciences, Scarborough, Ont., Canada M1C 1A4 b Institute for Brain Aging and Dementia, 1226 Gillespie Dr, University of California, Irvine, CA 92697-4540, USA c Science and Technology Center, Hill’s Pet Nutrition, Inc., P.O. Box 1658, Topeka, KS 66601-1658, USA d Lovelace Respiratory Research Institute, 2452 Ridgecrest Dr, S.E., Albuquerque, NM 87108, USA Received 26 September 2003; received in revised form 14 January 2004; accepted 17 February 2004 Abstract
The effectiveness of two interventions, dietary fortification with antioxidants and a program of behavioral enrichment, was assessed in a longitudinal study of cognitive aging in beagle dogs. A baseline protocol of cognitive testing was used to select four cognitively equivalentgroups: control food-control experience (C-C), control food-enriched experience (C-E), antioxidant fortified food-control experience (A-C),and antioxidant fortified food-enriched experience(A-E). We also included two groups of young behaviorally enriched dogs, one receivingthe control food and the other the fortified food. Discrimination learning and reversal was assessed after one year of treatment with a sizediscrimination task, and again after two years with a black/white discrimination task. The four aged groups were comparable at baseline.
At one and two years, the aged combined treatment group showed more accurate learning than the other aged groups. Discriminationlearning was significantly improved by behavioral enrichment. Reversal learning was improved by both behavioral enrichment and dietaryfortification. By contrast, the fortified food had no effect on the young dogs. These results suggest that behavioral enrichment or dietaryfortification with antioxidants over a long-duration can slow age-dependent cognitive decline, and that the two treatments together aremore effective than either alone in older dogs.
2004 Elsevier Inc. All rights reserved.
Keywords: Antioxidants; Beagle; Mitochondrial co-factors; Discrimination and reversal learning; Behavioral enrichment; Aging 1. Introduction
aged animals are compared with selected groups of youngeranimals. This strategy has limitations, which include cohort Over the past several years, our laboratories have been effects and the possibility of selective bias in mortality. Co- studying a novel model of cognitive aging, that of the aged hort effects may occur because of factors extrinsic from ag- beagle dog. We have previously established that dogs show ing, which could result in scores that are either particularly marked age-dependent decline in learning and memory, low or particularly high In primate aging studies, for example, aged animals often reside in the same lab for most pattern of cognitive decline mirrors that seen in humans in of their lives and have undergone considerable cognitive several respects Aged dogs also develop neuropathol- testing, which could enhance their cognitive performance, ogy that is similar to that seen in both successfully aging when compared to experimentally na¨ıve aged animals. Cog- humans and patients with Alzheimer’s disease. Like hu- nitive performance of aged dogs also is also enhanced by mans, beta amyloid protein is deposited in the aging dog brain and shows a selective brain distribution that The present study sought to further extend the canine model of cognitive aging with a two-year longitudinal in- To date, our analysis of canine cognitive aging has been re- vestigation of discrimination and reversal learning ability stricted to cross sectional studies, in which selected groups of in groups of young and aged beagle dogs. We had twomajor goals. The first was to obtain longitudinal data of ∗ Corresponding author. Tel.: +1-416-287-7402; fax: +1-416-287-7642.
age-dependent decline in learning ability. The second was E-mail address: milgram@psych.utoronto.ca (N.W. Milgram).
to assess the effectiveness of two interventions in counter- 0197-4580/$ – see front matter 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.neurobiolaging.2004.02.014 N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 acting age-dependent decline, behavioral enrichment and a and in the SAMP8 mouse In clinical trials, an- maintenance food fortified with a broad spectrum of antiox- tioxidant supplementation of Alzheimer’s patients with Vi- tamin E was found to delay the onset of institutionalizationWe have also found that short-term maintenance on 1.1. Behavioral enrichment and age-dependent cognitive an antioxidant fortified food improved discrimination learn- ing in beagle dogs we have previouslyshown evidence of increased oxidative stress in the aged ca- The behavioral enrichment intervention included three components: increased exercise, environmental enrichment, The present investigation started with a period of baseline and possibly most important, a program of cognitive enrich- testing, which was used to separate beagle dogs, approxi- ment. Exercise was suggested primarily by studies indicating mately 8–11 years of age, into four cognitively equivalent that physical activity is associated with improved cognitive groups, which differed in food provided and behavioral en- function and lower risks of cognitive impairment and de- richment. We hypothesized that both the behavioral enrich- mentia The second component, environmental en- ment and dietary supplementation treatments would have richment, was suggested by evidence that rearing in enriched beneficial effects on cognitive function, and that the two environments improves learning ability, produces beneficial treatments combined would be more effective than either by changes in cellular structure and increases the résistance of itself. To partially evaluate this hypothesis, one year after neurons to injury effect can be sufficiently robust as the treatment phase was initiated the animals were tested to reduce or eliminate age-dependent cognitive decline successively on a size discrimination learning task and a The rationale for including a cognitive enrichment inter- size discrimination reversal learning task. These tasks were vention was based on retrospective studies of human sub- selected because they were conceptually similar to the ob- jects suggesting a link between cognitive experience and the ject discrimination learning and object discrimination rever- development of age-dependent cognitive dysfunction. Peo- sal tasks used in assessing baseline cognitive function. Fur- ple characterized as having a low level of cognitive func- thermore, we have previously found the size discrimination tion are more likely to develop severe cognitive dysfunction and size discrimination reversal tasks to show age-sensitivity than people characterized as having a high level of cogni- The one-year results have been previously reported tive function , several studies have reported an Both the fortified food and behavioral enrichment im- inverse relationship between amount of education and rate proved learning, most notably in the reversal learning task.
of cognitive decline later in life. More direct evidence has However, these findings were mainly attributable to the su- been obtained in studies demonstrating that special training perior performance of the combined treatment group (forti- protocols improve cognitive performance in the elderly fied food and enriched experience), when compared to the other three groups. Approximately two years following thestart of the treatment phase, the animals were tested on a 1.2. Antioxidant supplementation and age-dependent black/white discrimination and a black/white discrimination reversal learning task, to provide a protocol for assessinglongitudinal changes in discrimination and reversal learning The dietary intervention consisted of providing a dry and the effectiveness of the fortified food and behavioral maintenance dog food fortified with a broad spectrum of enrichment interventions over two years.
antioxidants and mitochondrial cofactors. The food wasintended to reduce damage to tissue by reactive oxygenspecies as well as support mitochondrial function. Reac- 2. Materials and methods
tive oxygen species are formed as by-products of cellularmetabolism and, when produced in amounts in excess of detoxification, are purported to cause oxidative stress Oxidative damage to proteins and lipids has been linked Forty-eight old and 17 young beagle dogs were trained to the development and accumulation of neuropathology on a battery of cognitive tests over approximately 9 months.
associated with degenerative disease as well as normal ag- Performance on the baseline testing was then used to divide ing and is therefore a likely causal factor in the aged dogs into four cognitively equivalent test groups of 12 animals each, with two treatment conditions—dietary There is also more direct evidence of beneficial effects fortification and behavioral enrichment. Group, C-C, was of dietary supplementation with antioxidants on cognition.
both fed the control food and provided with control experi- In aged rats, antioxidants improved spatial learning ence; group, C-E, received the control food and a program motor learning and cerebellar function The effective- of enriched experience. group A-C dogs were fed food forti- ness of antioxidants in counteracting age-related cognitive fied with antioxidants and mitochondrial cofactors and also decline has also recently been demonstrated in mice defi- given control experience: while group A-E received both the cient in apolipoprotein Vitamin E deficient aged rats fortified food and the enriched experience.
N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 The young beagle dogs were divided into two cognitively adjusted so that the animals maintained a relatively constant equivalent groups, one of which was given the antioxidant body weight throughout the duration of the study.
enriched food (N = 9) and the other the control food (N = The aged subjects were all administered complete physi- 8). One control dog was subsequently dropped from the cal and neurological examinations prior to the dietary inter- study during the first treatment year because of motivational vention and again every six months after the start of inter- problems, reducing the size of the control group to seven.
vention. Dogs were also initially examined by slit-lamp for Both young groups received the behavioral enrichment pro- ocular abnormalities that might have impaired the animals’ visual capabilities. The initial physical examinations did not One year after the start of the dietary manipulation, all reveal neurological, musculoskeletal, ocular or physical ab- dogs were tested on both a size discrimination and a rever- normalities that justified exclusion from the study.
sal learning task. Prior to the one-year test, dogs in the en-riched environment groups had participated in a landmark discrimination task an oddity discrimination taskThe size discrimination task evaluated ability to learn The test apparatus was a 0.609 m × 1.15 m × 1.08 m to distinguish two objects that differed only in size in order wooden chamber that was based on a canine adaptation to locate a food reward. In the size discrimination reversal of the Wisconsin General Test Apparatus The testing task, the association between the objects and reward was chamber was equipped with a sliding Plexiglas food tray switched. Thus, if an animal was rewarded for approaching with three food wells. Vertical stainless steel bars covered the smaller of two objects during the initial discrimination the front of the box. The height of the bars was adjustable learning task, it was rewarded for approaching the larger of to allow the size of the opening to each food well to be the two objects during the reversal task. The size discrimi- uniquely adjusted for each dog. The experimenter was sepa- nation results have been previously reported rated visually from the dog, by a wooden screen containing The enrichment condition between years 1 and 2 of treat- a one-way mirror, and a hinged wooden door at the bottom.
ment consisted of training on a size concept learning task Testing was conducted in darkness, except for a light with on a repeated reversal learning task. At the com- a 60 W bulb attached to the front of the box. Each test trial pletion of the two-year enrichment phase, all animals were commenced with the hinged door being opened for the pre- trained on a black/white discrimination learning task, in sentation of the tray. A 1 cm3 amount of Hill’s® Prescription which the subjects were presented with two blocks that were Diet® Canine p/d(tm) canned food was used as the reward.
identical in size and shape but differed in color, with oneobject painted black and the other white. The animals were first trained to approach one of the two to obtain a food re-ward. After achieving a criterion level of performance, the The control and antioxidant foods were formulated to rewarded objects were then switched for the black/white dis- meet the nutrient profile for the American Association of Feed Control Officials recommendations for adult dogs(AAFCO 1999) two foods were identical, except for the inclusion of a broad-based antioxidant and mitochondrialcofactor supplementation to the test food. The control and Two groups of beagle dogs (Canis familiaris) served as enriched foods had the following differences in formulation subjects. The first consisted of 48 aged dogs (24 males and on an as fed basis, respectively: d,l-alpha-tocopherol ac- 24 females) ranging from 7.2 to 11.6 years at the start of etate (120 ppm versus approximately1000 ppm), l-carnitine baseline testing, and from 8.05 to 12.04 years of age at the (<20 ppm versus approximately 275 ppm), d,l-alpha-lipoic start of the treatment phase. The second group consisted of acid (<20 ppm versus approximately 125 ppm), ascorbic 17 young dogs (6 males and 11 females), 1.3–3.9 years of acid as Stay-C (<30 ppm versus approximately 80 ppm), age at the start of baseline testing testing and 1.95–4.6 at the and 1% inclusions of each of the following (1:1 exchange start of the treatment phase. Half the dogs (young and old) for corn): spinach flakes, tomato pomace, grape pomace, came from a closed colony of beagle dogs (Cohort 1). The carrot granules and citrus pulp. The rationale for these in- other half were obtained from a second, independent, closed clusions is as follows: Vitamin E is lipid soluble and acts colony (Cohort 2). Both groups were na¨ıve with respect to to protect cell membranes from oxidative damage; Vitamin cognitive test experience before starting the study.
C is essential in maintaining oxidative protection for the The aged dogs were housed, either singly or in pairs, in soluble phase of cells as well as preventing Vitamin E from pens with continual access to fresh water. Because of space propagating free radical production; alpha-lipoic acid is a considerations, the young dogs were housed in a separate cofactor for the mitochondrial respiratory chain enzymes, animal facility, from two to four per room. In all other re- pyruvate and alpha-ketoglutarate dehydrogenases, as well as spects, the old and young animals were treated identically.
an antioxidant capable of redox recycling other antioxidants All dogs were fed approximately 300 g of the control food and raising intracellular glutathione levels; l-carnitine is a once daily. In each case, however, the quantity of food was precursor to acetyl-l-carnitine and is involved in mitochon- N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 drial lipid metabolism and maintaining efficient function; fruits and vegetables are rich in flavonoids and carotenoids Testing on the size discrimination learning task com- and other antioxidants. The diet was produced by an ex- menced approximately 20 months following the start of trusion process and was fed for no more than six months baseline testing, and approximately one year after the start of the treatment phase. For the animals in the behaviorallyenriched group (group C-E and group A-E), a one-week 2.5. Behavioral enrichment intervention non-test interval preceded the start of size discriminationlearning. The dogs in the control experience (A-C, C-C) The behavioral enrichment condition commenced after condition did not undergo any cognitive testing for approx- completion of the baseline cognitive testing. The animals in imately nine months after completing the baseline testing.
the enriched group were housed with kennel mates, exercised The procedure followed in the size and size reversal learning twice a week for 15 min intervals, and given sets of toys that were alternated weekly. None of these were provided to thecontrol animals. The enrichment condition also included a cognitive enrichment protocol. The first year of cognitive en- Training on the black/white discrimination task started richment started immediately after baseline with testing on approximately two years after starting the treatment condi- a series of landmark discrimination problems, as previously tions. We used two wooden blocks that were identical in all described After completing the landmark task, the sub- respects except color: one was covered with white enamel jects were tested on to a series of oddity discrimination lean- paint and the other black enamel paint. The subjects were ing problems After completing the oddity problems, first administered a preference test, which consisted of a sin- the dogs were then tested for retention of the landmark task.
gle test session of 10 trials used to establish object prefer- The cognitive enrichment provided during the second year ences; the total choices of one the blocks out of 10 provided consisted of a series of nine discrimination learning tasks the absolute preference score. On this and all subsequent test that were intended to study size concept learning a sessions, the objects were placed over the two lateral food series of repeated reversal learning tasks that were intended wells, and the location of the objects varied randomly, with the constraint that each object was placed on each lateralfood well on exactly 50% of the trials. A customized com- puter program controlled all timing and randomization pro-cedures. The program also assured that on each trial, the lo- cations of the objects were the same for each animal. Before All subjects underwent a standard pretraining cognitive the beginning of each trial, the computer emitted a tone that testing protocol. It consisted of reward approach and ob- served as a cue for the dog and instructed the experimenter ject approach learning were procedural learning to present the food tray. Each trial was started when the ex- tasks designed to train animals to displace an object on a tray perimenter pressed a key and simultaneously presented the to obtain a food reward consisting of approximately 1 g of tray to the subject. The dogs’ responses to the two objects Hill’s® Prescription Diet® Canine p/dTM canned food. The were recorded by a key press, which also indicated the end dogs responded to the objects by pushing them away from of the trial and signalled the beginning of the inter-trial in- the food well with their noses, and then eating the food.
After completing the procedural learning tasks, all subjects Training on the black/white discrimination problem were trained on an object discrimination learning task, which started on the day following the preference test. The ani- was followed by an object reversal learning task The mals were given 10 trials per day, constituting one session, animals were then tested on an object recognition memory with an intertrial interval of 30 s. Testing was six days per task delayed-non-matching-to-position task (DNMP) week. The animals received a maximum of 40 training The initial group assignment took into consideration sessions to achieve a two-stage criterion. The first stage age, sex, cohort and the subjects combined performance on was successfully met once the animal either averaged 80% the reversal learning task, the object recognition task, and over two sessions, or at least 90% on a single session. To the DNMP task. The baseline cognitive data has previously complete the second stage, the dog was required to respond been reported four test groups of aged dogs did not correctly on at least 70% of the trials over three successive differ on any of the baseline tests used in classification. Sim- sessions. Thus, passing both stages took a minimum of four ilarly, the two test groups of young animals did differ from test sessions. One dog failed to learn the black/white dis- each other on the baseline evaluations. There were, how- crimination task within the 40 trials, and was consequently ever, significant differences between the old and young ani- administered a program of remedial training, so that the mals in performance on the reversal and visuospatial tasks.
dog could be tested on the reversal task. During the reme- All animals were maintained on the control food during the dial training phase, an additional 13 training sessions were pretraining period, which continued for approximately nine allotted for each animal to reach the criterion performance N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 The black/white reversal task started on the day following 3.2. Effect of food and experience on learning a completion of the black/white discrimination learning task.
black/white discrimination and reversal task The testing procedures were identical, except that the rewardcontingencies of the positive and negative block were re- For the aged animals, the results of the black/white task versed. Thus, if an animal was rewarded for approaching the were first analyzed with a repeated measures analysis of white block during the initial testing, it was now rewarded variance, with discrimination and reversal learning as within subject measures, and cohort, food, and behavioral enrich-ment as between subject measures. The results revealed sig- nificant main effects of food (F(1, 34) = 4.678; P < 0.05),behavioral enrichment (F(1, 34) = 31.89, P < 0.01) and For individual subjects, rate of learning was character- task (F(1, 34) = 78.93; P < 0.01). As expected, the task ized by error scores, which were calculated by adding the effect was due to the reversal task being more difficult than total number of errors to either pass the two-stage learn- the original discrimination learning task.
ing criterion, or the total number of errors made over To further breakdown the food and behavioral enrich- 40 training trials. The data were analyzed with factorial ment effects, we performed separate factorial analyses for and repeated measures analysis of variance (ANOVA) the black/white discrimination task and for the black/white When required, post hoc analysis was performed by discrimination reversal task. On the black/white discrimina- Tukey’s studentized range test (HSD), using the 0.05 level tion task, the ANOVA revealed a significant main effect of of significance. In addition, chi-square analysis was per- behavioral enrichment (F(1, 38) = 22.35; P < 0.01), and formed on the frequency of failure for the three assessment no other significant effects or interactions. illus- trates that each of the groups that received behavioral en- The initial analyses took into consideration food, behav- richment (group A-E and group C-E) made fewer errors than ioral enrichment condition, cohort and sex. There were no either of the two non-enriched groups (group A-C and group significant effects of sex on any of the dependent variables C-C). Tukey’s LSDs indicated that each behavioral enrich- and sex was therefore dropped from all subsequent analyses.
ment groups performed significantly better than its respec-tive control enrichment group (group C-C versus group C-Eand group A-C versus group A-E). By contrast, the control 3. Results
groups did not differ from each other.
Analysis of black/white discrimination reversal learning, on the other hand, revealed significant effects of both behav-ioral enrichment (F(1, 34) = 27.2094; P < 0.01) and food shows the sample size and mean ages of each (F(1, 34) = 5.11; P < 0.05). These results are largely due group at the start of discrimination testing during baseline, to the high level of performance of group A-E (see after one year of treatment and after two years. As indi- which received the combined treatment of antioxidant en- cated in completed sets of discrimination and re- riched food and behavioral enrichment. Further, post hoc versal data were not obtained from four animals assigned analysis revealed that group A-E did significantly better than to group C-C, and 1 animal assigned to group C-E: Four of group C-C and group A-C. The only other significant group these died or had to be euthanized for medical reasons; the differences were between group C-E and group C-C.
fifth was dropped from the study because of motivational We also analyzed the data from the young animals using a repeated measures ANOVA, and found a significant effect of Table 1Mean age of groups at start of discrimination testing over three years Mean ages and standard deviations of the four groups of old dogs and two groups of young dogs at the start of baseline discrimination learning testing,after one year of treatment, and after two years of treatment.
N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 Fig. 2. Effect of antioxidant fortified food on acquisition of a black/white discrimination and black/white discrimination reversal learning task inyoung beagle dogs. Error bars represent standard errors.
There were no significant differences in the black/white discrimination learning task. On the black/white discrimina-tion reversal task, by contrast, there was a highly significant effect of age (F(1, 34) = 17.12; P < 0.01). As indicated in this was largely due to poor performance of the oldanimals on the control food. Post hoc analysis indicated this group performed significantly more poorly than both groups Fig. 1. Effect of antioxidant fortified food and behavioral enrichment onacquisition of a black/white discrimination learning task (A) and reversal learning task (B) in aged beagle dogs. Error bars represent standard errors(C-C: control food-control experience, C-E: control food-enriched experi- ence, A-C: antioxidant fortified food-control experience, A-E: antioxidant fortified food-enriched experience). Bars with different superscripts are significantly different by Tukey’s studentized range test (HSD).
task (F(1, 12) = 28.17; P < 0.01), but no other significant Because the young animals all received behavioral en- richment, age differences on both the black/white discrim- ination and reversal tasks were assessed with the use of a repeated measures ANOVA that compared the two younggroups with the old groups provided with behavioral enrich- ment. In this analysis, task (discrimination versus reversallearning) served as the within subject variable and food and age were between subject variables. The results revealed a highly significant effect of task (F(1, 34) = 48.93; P < 0.01), age (F(1, 34) = 15.26; P < 0.01), and a significantage by task interaction. (F(1, 34) = 5.799; P ≤ 0.03). To Fig. 3. Age differences in acquisition of black/white discrimination andblack/white discrimination reversal learning task. The old animal data further clarify these results, we carried out separate factorial was taken only from subjects in the behaviorally enriched groups. Error analysis for the black/white discrimination and black/white bars represent standard errors. Groups with different superscripts are significantly different by Tukey’s studentized range test (HSD).
N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 of young animals and than the old animals on the fortified (F(2, 66) = 34.11, P < 0.01). ws that the task effect reflects the animals showing a progressive slowingin learning over the three years. The interaction with expe- 3.4. Longitudinal changes in discrimination and reversal rience reflects the behaviorally enriched group performing learning between baseline and year 2 assessment better than the non-enriched group over the third year only.
Finally, the task by food effect reflects consistently better 3.4.1. Overall results with all animals, young and old performance of the enriched animals over the controls on the last two years, after the start of the antioxidant treat- To evaluate longitudinal changes in discrimination and reversal learning, the data over the three years were first an- The reversal learning was first analyzed with a repeated alyzed with an omnibus repeated measures ANOVA with measure ANOVA over the three years and revealed signif- age (young versus old) and food (enriched versus control) icant main effects of behavioral enrichment (F(1, 34) = as between subject variables and test year (object, size and 9.78 P < 0.001), food (F(1, 34) = 4.198, P < 0.05), black/white) and task type (discrimination versus reversal) and task (F(2, 68) = 52.11, P ≤ 0.001). There were as within subject factors. There were highly significant main also significant interactions between task and enrichment effects of age (F(1, 54) = 44.277; P < 0.0001), test year (F(2, 68) = 15.60, P ≤ 0.0001) and between task and (F(2, 108) = 21.149) and task type (F(1, 54) = 273.67; cohort (F(2, 68) = 3.37, P < 0.05). The behavioral en- P < 0.0001). There were also significant interactions be- richment effect is shown in which illustrates that tween age and test year (F(2, 108) = 10.446; P ≤ 0.001) the animals provided with the behavioral enrichment treat- and age and task type (F(1, 54) = 24.76; P ≤ 0.00001).
ment learned more accurately than the animals providedwith control experience, and the differences increased over 3.4.2. Effects of age within behaviorally enriched groups repeated testing. illustrates that the food effect is To examine the effects of age, we next performed the same due to improved performance of the group given both the analysis on the behaviorally enriched dogs only, which in- antioxidant fortified food in the second treatment year on cluded all of the young dogs and half of the old dogs. The re- sults revealed a highly significant effects of food (F(1, 34) = The subjects tested in these experiments were allowed a 7.22; P ≤ 0.02), test year (F(2, 68) = 15.34; P < 0.0001) maximum of 40 sessions to solve the reversal learning task, and task type (F(1, 34) = 142.04; P < 0.0001). There and some of the animals were unsuccessful. The error score were also significant interactions between age and test year assigned to the subjects that failed was based on the 40 test (F(2, 68) = 7.75; P ≤ 0.001) and between age and task sessions administered, which underestimated the true error type (P(1, 34) = 18.9153; P < 0.001). As illustrated in rate because of a ceiling effect. The number of failures for 4, the task and age effects are attributable to animals gen- each of the test groups is shown in The contrast erally showing faster learning of the black/white discrimi- between the animals given the combined treatment and the nation task than of the black/white discrimination reversal animals in the control-control group was notable. All 12 of task, and second, consistently more accurate learning by the the animals in the combined treatment condition were able to solve all of the reversal learning problems, while onlytwo of eight control-control animals showed learning. The 3.4.3. Effects of food and experience on aged animals chi-squared value obtained by comparing frequency of fail- We then looked at treatment effects in the aged ani- ure for the discrimination reversal learning over all time pe- mals alone, looking at discrimination and discrimination riods with expected frequencies was highly significant (P = reversal learning separately. For the discrimination learn- 0.02). Subsequent chi-square analysis of the failure at each ing task, there were highly significant main effects of task measured point showed no significant difference at base- (F(2, 68) = 34.11, P < 0.0001) and behavioral enrich- line or the first year of assessment. However, the second ment (F(1, 34) = 17.95, P < 0.001) and significant two year of assessment revealed a highly significant chi-square way interactions between task and behavioral enrichment of 0.0033. This could be attributed to the high failure rate (F(2, 66) = 7.475, P < 0.001) and between task and food in the C-C group compared to the other groups.
Table 2Reversal learning failures as a function of task and treatment group N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 Fig. 4. Longitudinal changes in discrimination (A) and reversal learning (B) in young and old beagle dogs. The aged group included in the figure waslimited to the subjects in the behaviorally enriched groups.
4. Discussion
task after two years receiving one of four treatment condi-tions: control food-control experience; fortified food-control This project had three goals: to evaluate the cognitive experience; control food-enriched experience; and fortified effectiveness of long-term maintenance on an antioxidant- food-enriched experience. We also tested two groups of dogs fortified food; to evaluate the effectiveness of a long-term that were young at the start of the study. One group received program of behavioral enrichment, and; to assess cogni- the control food and the other the fortified food. Both groups tive decline in the beagle dog in a longitudinal study. The of young dogs received behavioral enrichment.
data presented in this report was obtained from aged dogs Both treatment conditions improved performance of the tested on a black/white discrimination learning and reversal aged group. However, the effectiveness of the various treat- N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 Fig. 5. Performance on discrimination learning tasks as a function of antioxidant fortification and behavioral enrichment in aged dogs. (A) Learningaccuracy over three years for the animals in the enriched experience and control experience groups. (B) Comparison of the aged animals on the antioxidantfortified and control foods at baseline and over the next two years. As indicated in the figure captions, the dogs were tested on an object discriminationtask at baseline, a size discrimination task after one treatment year and a black/white discrimination task after two treatment years.
ment combinations varied as a function of both experience By contrast, all 12 animals in the combined treatment con- and food. Performance of the black/white discrimination dition successfully achieved the a priori learning criterion.
learning task was significantly improved in the aged dogs Finally, on both tasks, the two treatments combined were provided with behavioral enrichment, relative to the control more effective than either treatment alone.
enrichment condition. The reversal task, by contrast, was The performance of the subjects in the young group by significantly affected by both experience and food. Further- contrast, was unaffected by the dietary manipulation, which more, because the experimental design limited the number was not unexpected. The effectiveness of the food is theoret- of the training trials, the magnitude of the treatment effects ically linked to its’ ability to arrest or reverse cellular dys- was likely underestimated. Eight of 18 animals in the control function produced by excessive free radicals and imiprove- food condition failed the reversal test in the allotted 40 ses- ment of aged mitochondrial function. However, free-radical sions and 10 of 20 animals in the control experience failed.
based brain damage is minimal in younger animals N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 Fig. 6. Performance on discrimination reversal learning tasks as a function of antioxidant fortification and behavioral enrichment over three years in agedsubjects. (A) The scores over three years for the animals in the enriched experience and control experience groups. Group differences are apparent inthe first test year, after one year of treatment. (B) Comparison of the aged animals on the antioxidant fortified and control foods at baseline and overthe next two years.
The results of this study extend our previous report on one year. We have also found, however, that dietary forti- the effects of the antioxidant fortification and behavioral en- fication has significant beneficial effects after a short time richment on size discrimination and reversal learning, which frame among animals provided with behavioral enrichment was carried out after one year of treatment. Whereas the present study revealed a significant main effect of food byitself on the reversal learning, the one-year results indicated 4.1. Effects of behavioral enrichment that antioxidant supplementation was only effective when itwas combined with behavioral enrichment present The behavioral enrichment condition included a program results indicate that the effect of fortified food on cognition of cognitive enrichment, increased physical activity and en- is more robust after two years on the food than after only vironmental enrichment. The present results do not allow N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 us to distinguish the relative importance of each of these provement, and also were able to erase other age-dependent interventions. We suspect, however, that the cognitive en- richment was particularly important. First, previous animalstudies in which aged subjects are provided with environ- 4.3. Cross sectional age differences mental enrichment have had small inconsistent effects oncognition By contrast, training animals on particular The analysis of age-dependent cognitive decline was tasks early in life can produce long-lasting changes in the partially confounded by the absence of a non-enriched animals’ abilities to learn those tasks later in life young-animal group. The young animals, consequently, Cognitive enrichment protocols have also been found to could only be compared with the behaviorally enriched aged produce beneficial effects in elderly human subjects. Ball animals. The results revealed significant differences be- et al. xamined the effect of three distinct cognitive in- tween the age groups on the reversal learning, but not on the terventions (memory training, reasoning training, and speed discrimination learning. The absence of an age-dependent of processing training). They were all effective, but the ef- difference in discrimination learning contrasts with data ob- fectiveness was selective, and improved only the targeted tained from these animals after only one year of treatment cognitive ability. These results from human subjects suggest with other studies showing age-dependent deficits that cognitive enrichment protocols produce task-specific in complex discrimination learning tasks e attribute improvement. The present results are consistent with this these results to two factors: the first is the extensive test ex- suggestion. Although the cognitive enrichment protocol con- perience given to our behaviorally enriched aged animals.
sisted of a broad range of tasks, they could all be solved with The second relates to the age range of the young group. Al- a discrimination learning strategy. This was also true of the though we have used the term young, the mean age of the black/white discrimination task, and the discrimination re- young group at the time of final testing was greater than versal task, which requires two cognitive skills: learning to five and some of the animals were over seven, which we inhibit the response to a previously rewarded stimulus and have previously considered to be middle aged learning to respond to a previously non-rewarded stimulus.
We did get significant age differences in the reversal learn- ing task, as expected based on our previous work 4.2. Effects of dietary intervention studies with non-human primates owever, thedifferences were largely a result of poorer performance of The finding of improved performance in the groups on the aged group that received the control food, suggesting the antioxidant fortified food is consistent with our previous that the antioxidant food was able to reduce age-dependent reports, which were obtained in animals provided with the fortified food for under a year The present resultsextend these findings to indicate that the fortified food re- 4.4. Longitudinal changes in cognitive ability in the young mains an effective therapeutic after two years of treatment.
The dietary intervention used in this study has been previ-ously described and discussed , the food was The other aim of this study was to obtain longitudinal enriched with a cocktail containing both antioxidants and mi- evidence of changes in discrimination and reversal learn- tochondrial cofactors. Because of the numerous ingredients, ing ability in the beagle dog. To characterize cognitive de- we do not know which if any specific component is partic- cline, we studied two groups of beagle dogs, an aged group ularly important, or, alternatively, whether beneficial effects and a young group. The aged group consisted of 48 dogs depend on the use of a broad spectrum of ingredients. The that ranged from approximately 8–10 years of age. The latter interpretation is consistent with the moderately large young group consisted of 17 dogs. The results demonstrated literature on the effects of antioxidant supplementation on progressive deterioration in performance over three years cognition. In several studies, in which only a limited group of in both discrimination and reversal learning in the aged antioxidants were used, the effects on cognition were small and restricted. Thus, Sano et al. that supplemen- At the start of the study, the old group performed signif- tation with ␣-tocopherol over two years did not improve icantly worse than the young on object discrimination re- scores on cognitive measures, such as on the mini-mental versal learning, but not on object discrimination learning, state examination, in moderately demented Alzheimer’s dis- suggesting overall cognitive impairment in 8- to 10-year-old ease patients. Socci et al. no effect on passive beagle dogs. As a group, the aged animals showed progres- avoidance memory of long-term antioxidant supplementa- sively poorer performance throughout the course of the study tion with Vitamin E, phenyl-␣-tert-butylnitrone and ascorbic on both discrimination and reversal tasks. The performance acid, although the antioxidant treatment did improve rate of differences between the size and black/white tasks are likely water maze learning. Finally, Joseph et al. not find due to marked age-dependent decline over the course of the differences between control treatment and a variety of an- study, rather than to differences in task difficulty. In a pre- tioxidants on water maze learning, although they reported vious experiment, we used a crossover design to compare evidence suggesting that supplements resulted in greater im- acquisition of a size discrimination and black/white discrim- N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 ination tasks and found that the size task was significantly This suggested time frame for cognitive aging in the bea- more difficult than the black/white task for aged beagle dogs gle dog is consistent with imaging studies of aged dogs and In the present study, by contrast, the aged dogs did time course studies of beta amyloid deposition. Su et al. more poorly on the black/white discrimination, which was found that after the age of 10, ventricular volume increases tested in the second year of the study, than they did on exponentially. Head et al. found that neuropathogy, the size discrimination task, which was tested in the first manifested by the occurrence of beta amyloid deposition, year of the study. Since this is the opposite of what we first appears when the beagle dog at about eight years, when have seen previously, it’s likely that the poorer performance it is most prominent in the prefrontal cortex. At this time, reflects age differences (the animals were one year older) many cognitive functions are unimpaired, although deficits rather than differences in task difficulty. This conclusion is are manifested in functions, such as reversal learning, that further reinforced by the data from the young dogs, which are likely frontal lobe dependent. After 10 years of age, beta performed comparably on the black/white, size and size dis- amyloid accumulation is also notable in entorhinal, parietal, crimination tasks, as well as on the discrimination reversal and occipital cortices, which is also the time frame when more severe and widespread cognitive dysfunction occurs.
Another notable observation was the high incidence of The suggestion of an increasing proportion of subjects failures with increased age—particularly by the control severe cognitive decline after about 12 is also consistent group on the reversal learning task. Because these tasks with observational data obtained from studies of pet dogs, are not particularly difficult for young or middle aged in which cognitive impairment characterized by disorienta- beagle dogs, learning failures provide strong evidence of tion, disturbances in social interactions, impairment in house age-associated cognitive decline or dementia. The present training and disruption of sleep–wake cycles shows an in- results, therefore, suggest that the likelihood of cognitive creasing prevalence with advanced age. Thirty percent of decline increases precipitously after the age of 10. This 11- to 12-year-old animals show impairment in one or more conclusion is also consistent with the results of Patronek category and 10% show impairments in two or more of these et al. in which a 10-year-old beagle was found to be categories. In animals between 15 and 16, the percentages equivalent in physiological age to a 66.6-year-old human.
We have also found evidence of cognitive deterioration inthe performance of dogs used in this study on a delayednon-matching to position task, a measure of visuospatial 5. Conclusions
function, which will be reported separately. Several ageddogs failed the task in the third year of the study, despite To conclude, the present results demonstrates that both the fact that they had successfully learned the task when discrimination and reversal learning ability decline progres- sively with advanced age in beagle dogs, but that the rate of We have focused primarily on longitudinal changes in the decline can be delayed by both behavioral enrichment and old dogs. We also found evidence of cognitive deterioration antioxidant supplementation. Possibly the most important in the young dogs. The performance of the young group outcome of this study was the demonstration that the behav- showed an overall, but not significant, decline in the second ioral enrichment and the antioxidant supplementation con- year of the study (on black/white discrimination and rever- dition combined were more effective than either alone. This sal), when compared to the first (size discrimination and re- study is the first that we know of to look at both interven- versal). However, we expected the dogs to do better on the tions in combination. The dietary intervention was based on black/white test than on the size discrimination based on a cocktail of compounds, and it will be important in future previous findings In fact, by the second treatment year studies to determine which of the ingredients are most effec- of the study, the designation of the group as young dogs was tive, and whether the cocktail can be improved. The behav- no longer appropriate, as the mean age of the dogs was now ioral intervention also involved a cocktail of treatments (ac- over five years. One possibility is that the performance on tivity, environmental enrichment and cognitive enrichment), the black/white discrimination task actually represents im- and the contribution of each to the present results remains paired performance, relative to younger dogs.
Collectively, these results suggest a possible time frame for the development of cognitive deterioration in the bea-gle dog. The data from the young dogs suggests that max-imal performance is typically reached between about two 6. Conflict of interest statement
and four years of age, and that performance begins to falloff around five years of age. By eight years of age, there The following conflict of interest was declared by the are clear and consistent age-dependent impairments. Subse- authors with respect to publication of this paper: Steven quently, cognitive decline increases at an accelerated rate, Zicker is an employee of Hill’s Pet Nutrition Inc., which and after about 12, an increasing proportion of animals show has commercialized the antioxidant fortified food used in severe decline and can be characterized as demented.
N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 Acknowledgments
[17] Fukui K, Omoi NO, Hayasaka T, Shinnkai T, Suzuki S, Abe K, et al. Cognitive impairment of rats caused by oxidative stress and agingand its prevention by Vitamin E. Ann NY Acad Sci 2002;959:275– This project was sponsored by funds provided by the National Institute of Aging (Grant AG12694) and [18] Gabbita SP, Lovell MA, Markesbery WR. Increased nuclear DNA by the U.S. Department of the Army, Contract No.
oxidation in the brain in Alzhemier’s disease. J Neurochem [19] Giaccone G, Verga L, Finazzi M, Pollo B, Taglaivini F, Frangione B, et al. Cerebral preamyloid deposits and congophilic angiopathyin aged dogs. Neurosci Lett 1990;114:178–83.
References
[20] Greiner PA, Snowdon DA, Schmitt FA. The loss of independence in activities of daily living: the role of low normal cognitive function [1] Adams B, Chan ADF, Callahan H, Siwak CT, Tapp D, Ikeda-Douglas in elderly nuns. Am J Pub Health 1996;86(1):62–6.
C, et al. Spatial learning and memory in the dog as a model of [21] Head E, Callahan H, Muggenburg BA, Cotman CW, Milgram NW.
cognitive aging. Behav Brain Res 2000;108:47–56.
Visual-discrimination learning ability and beta-amyloid accumulation [2] Adams B, Chan A, Callahan H, Milgram NW. The canine in the dog. Neurobiol Aging 1998;19(5):415–25.
as a model of aging and dementia: recent developments. Prog [22] Head E, McCleary R, Hahn FF, Milgram NW, Cotman CW.
Neuro-Psychopharmacol Biol Psychiatry 2000;5:675–92.
Region-specific age at onset of ␤-amyloid in dogs. Neurobiol Aging [3] Aksenova MV, Aksenova MY, Payne RM, Trojanowski JQ, Schmidt ML, Carney JM, et al. Oxidation of cytosolic proteins and expression [23] Head E, Liu J, Hagen TM, Muggenburg BA, Milgram NW, Ames of creatine kinase B in frontal lobe in different neurodegenerative BN, et al. Oxidative damage increases with age in a canine model disorders. Dement Geriatr Cogn Disord 1999;10:158–65.
of human brain aging. J Neurochem 2002;82(2):375–81.
[4] Association of American Feed Control Officials: AAFCO dog and cat [24] Ikeda-Douglas CJ, Zicker SC, Estrada J, Jewell DE, Milgram NW.
food nutrient profiles. In: 1999 Official Publication of American Feed Prior experience, antioxidants and mitochondrial cofactors improve Control Officials Incorporate. West Lafayette, IN: AAFCO;1999.
cognitive function in aged beagles. Vet Ther 2004;5(1):5–16.
[25] Joseph JA, Shukitt-Hale B, Denisova NA, Bielinski D, Martin A, [5] Ball K, Berch DB, Helmer KF, Jobe JB, Leveck MD, Marsiske M, McEwen JJ, et al. Reversals of age-related declines in neuronal et al. Effect of cognitive training interventions with older adults.
signal transduction, cognitive, and motor behavioral deficits with blueberry, spinach or strawberry dietary supplementation. J Neurosci1999;19:8114–21.
[6] Bartus RT, Dean RL, Fleming DL. Aging in the rhesus monkey: [26] Lai ZC, Moss MB, Killiany RJ, Rosene DL, Herndon JG. Executive effects on visual discrimination learning and reversal learning. J system dysfunction in the aged monkey: spatial and object reversal learning. Neurobiol Aging 1995;16:947–54.
[7] Beckman KB, Ames BN. The free radical theory of aging matures.
[27] Markesbery WR, Lovell MA. Four-hydroxynonenal, a product of lipid peroxidation is increased in the brain of Alzheimer’s disease.
[8] Bickford PC, Gould T, Briederick L, Chadman K, Pollock A, Young D, et al. Antioxidant-rich diets improve cerebellar physiology and [28] Mattson MP, Duan W, Lee J, Guo Z. Suppression of brain motor learning in aged rats. Brain Res 2000;866:211–7.
aging and neurodegenerative disorders by dietary restriction and [9] Callahan H, Ikeda-Douglas C, Head E, Cotman CW, Milgram environmental enrichment: molecular mechanisms. Mech Ageing Dev NW. Development of a protocol for studying object recognition memory in the dog. Prog Neuro-Psychopharmacol Biol Psychiatry [29] Milgram NW, Head E, Weiner E, Thomas E. Cognitive functions and aging in the dog: acquisition of non spatial visual tasks. Behav [10] Cartford MC, Gemma C, Bickford PC. Eighteen-month-old Fischer 344 rats fed a spinach-enriched diet show improved delay classical [30] Milgram NW, Siwak CT, Gruet P, Atkinson P, Woehrlé F, Callhan eyeblink conditioning and reduced expression of tumor necrosis H. Oral administration of adrafinil improves discrimination learning factor alpha (TNFalpha) and TNFbeta in the cerebellum. J Neurosci in aged beagle dogs. Pharmacol Biochem Behav 2000;66:301–5.
[31] Milgram NW, Head E, Muggenburg B, Holowachuk D, Murphey [11] Chan ADF, Nippak P, Murphey H, Ikeda-Douglas C, Muggenberg B, H, Estrada CJ, et al. Landmark discrimination learning in the dog: Head E, et al. Visuospatial impairments in aged canines: the role of effects of age, an antioxidant fortified diet, and cognitive strategy.
cognitive-behavioral flexibility. Behav Neurosci 2002;116:443–54.
Neurosci Biobehav Rev 2002;26:679–95.
[12] Churchill JD, Galvez R, Colcombe S, Swain RA, Kramer AF, [32] Milgram NW, Zicker SC, Head E, Muggenburg BA, Murphey H, Greenough WT. Exercise, experience and the aging brain. Neurobiol Ikeda-Douglas C, et al. Dietary enrichment counteracts age-associated cognitive dysfunction in canines. Neurobiol Aging 2002;23:737–45.
[13] Cotman CW, Berchtold NC. Exercise: a behavioral intervention to [33] Milgram NW. Cognitive experience and its effect on age-dependent enhance brain health and plasticity. Trends Neurosci 2002;25:295– cognitive decline in beagle dogs. Neurochem Res 2003;28:1677–82.
[34] Milgram NW, Head E, Zicker S, Ikeda-Douglas CJ, Murphey H, [14] Cummings BJ, Head E, Ruehl W, Milgram NW, Cotman CW. The Muggenburg B, et al. Dietary antioxidant fortification and behavioural canine as an animal model of human aging and dementia. Neurobiol enrichment combined improve cognitive performance of aged beagles on visual discrimination learning and reversal, Expt Gerontol 2004, [15] Escorihuela RM, Tobena A, Fernandez-Teruel A. Environmental enrichment and postnatal handling prevent spatial learning deficits [35] Miranda S, Opazo C, Larrondo LF, Munoz FJ, Ruiz F, Leighton F, in aged hypoemotional (roman high-avoidance) and hyperemotional et al. The role of oxidative stress in the toxicity induced by amyloid (roman low-avoidance) rats. Learn Mem 1995;2:40–8.
beta-peptide in Alzheimer’s disease. Prog Neurobiol 2000;62:633– [16] Farr SA, Poon HF, Dogrukol-Ak D, Drake J, Banks WA, Eyerman E, et al. The antioxidants alpha-lipoic acid and N-acetylcysteine reverse [36] Moore S, Sandman CA, McGrady K, Kesslak JP. Memory training memory impairment and brain oxidative stress in aged SAMP8 mice.
improves cognitive ability in patients with dementia. Neuropsychol N.W. Milgram et al. / Neurobiology of Aging 26 (2005) 77–90 [37] Neilson JC, Hart BL, Cliff KD, Ruehl WW. Prevalence of behavioural [45] Tapp PD, Siwak CT, Estrada J, Muggenburg BA, Head E, Cotman changes associated with age-related cognitive impairment in dogs.
CW, et al. Size and reversal learning in the beagle dog as a measure of executive function and inhibitory control in aging. Learn Mem [38] Patronek GJ, Waters DJ, Glickman LT. Comparative longevity of pet dogs and humans: implications for gerontology research. J Gerontol [46] Tapp PD, Siwak CT, Head E, Cotman CW, Murphey H, Muggenberg A Biol Sci Med Sci 1997;52(3):B171–8.
BA, et al. Concept abstraction in the dog: development of a protocol [39] Rapp PR. Visual discrimination and reversal learning in the aged using successive discrimination and size concept tests. Behav Brain monkey (Macaca mulatta). Behav Neurosci 1990;104:876–84.
[40] Rowe JW, Kahn RL. Human aging: usual and successful. Science [47] Van Gool WA, Mirmiran M, van Haaren F. Spatial memory and visual evoked potentials in young and old rats after housing in an [41] Sano M, Ernesto C, Thomas RG, Klauber MR, Schafer K, Grundman enriched environment. Behav Neural Biol 1985;44:454–69.
M, et al. A controlled trial of selegiline, alpha-tocopherol, or both [48] Veinbergs I, Mallory M, Sagara Y, Masliah E. Vitamin E as treatment for Alzheimer’s disease. The Alzheimer’s Disease supplementation prevents spatial learning deficits and dendritic Cooperative Study. N Engl J Med 1997;336(17):1216–22.
alterations in aged apolipoprotein E-deficient mice. Eur J Neurosci [42] SAS/STAT® users guide, version 6, vol. 2. 4th ed. Cary NC: SAS [49] Vicens P, Redolat R, Carrasco MC. Effects of early spatial training on [43] Socci DJ, Crandall BM, Arendash GW. Chronic antioxidant treatment water maze performance: a longitudinal study of mice. Exp Gerontol improves the cognitive performance of aged rats. Brain Res [50] Voytko ML. Impairments in acquisition and reversals of two-choice [44] Su M-Y, Head E, Brooks WM, Wang Z, Muggenberg BA, Adam GE, et al. MR imaging of anatomicand vascular characteristics in a canine model of human aging. Neurobiol Aging 1998;19:479–85.

Source: http://muskrat.middlebury.edu/lt/cr/faculty/shalpern-lt/memory/aged%20canines%20antiox%20and%20envir%20enrich%20Milgram%20et%20al%20NBA%202005.pdf