Science Watch

It's well known that memory tends to slacken with age, but some forms of it stay stronger longer, a new study suggests. Specifically, the study suggests that older people may retain the capacity for emotional learning longer than they retain the capacity for factual learning. If extended, the findings could support the use of revamped tests to distinguish among types of memory loss in diseases such as Alzheimer's and encourage therapists to use (positive) emotional props to foster learning.

The finding comes from one of the first systematic studies of aversive (fear) conditioning and awareness across the adult life span, led by Duke University Medical Center's Kevin LaBar, PhD, a cognitive neuroscientist, and Kathleen Welsh-Bohmer, PhD, a neuropsychologist specializing in aging and Alzheimer's disease. Using fear conditioning as a model, they studied 87 adults and found that the rate at which people learn in this fundamental way appears to be the same in young, middle-aged and elderly adults.

Observers say the findings, which appear in the October issue of Behavioral Neuroscience (Vol. 118, No. 5), move the field forward.

"Evidence based on [age-related] fear conditioning is sparse, and the study has made an important contribution to fill this gap," note Christian Bellebaum, PhD, and Irene Daum, PhD, cognitive neuroscientists at Germany's Ruhr-University of Bochum, who contributed a theoretical comment on it in the same journal.

Squaring off in the lab

In aversive (fear/anxiety) conditioning, experimenters link an innocuous conditioned stimulus (CS) to a distressing unconditioned stimulus (US). The Duke team tested the hypothesis that this type of emotional learning lasts better over time than, say, memory for the capitals of all 50 U.S. states. To see if the capacity for fear conditioning lasts a lifetime, the Duke researchers tested fear conditioning in three groups of adults: 46 adults ages 18 to 29 years, 17 adults ages 51 to 64 years, and 24 adults ages 66 to 80 years. None showed signs of probable Alzheimer's disease.

Participants looked at a computer screen that displayed red or green squares--a CS. Sometimes they heard bursts of fairly loud white noise (this US was, at 100 to 105 decibels, considerably louder than normal conversation but not enough to cause sensorineural hearing loss). Participants who then showed anxiety when they saw the noise-linked colored squares--via increased electrical conductance on their sweaty palms--were said to be aversively conditioned.

LaBar says the measurement makes sense: Anxiety revs up the sympathetic nervous system, causing people to sweat more, which raises conductance.

"The individual is processing the conditioned stimulus as a danger or warning signal to prepare the body's defenses," LaBar explains.

Meanwhile, learners may or may not be aware that they are actually learning, as the process can be explicit (creating declarative memory) or implicit (generating procedural memory). To control for this potential complication, the Duke team asked participants after each session what they knew about the relationship between stimuli. Participants were "aware" if they reported that a given colored square predicted noise. And, as expected, a greater proportion of elderly participants were unaware of the CS-US contingency due to the age-related drop in declarative memory.

The researchers note a key finding: When the data were combined, it seemed that as a group, older people didn't appear to get anxious when seeing the squares that had been linked to loud noise. But grouped data can mask the truth. When the researchers controlled for (the lessened) awareness by dividing participants in each age group into "aware" and "unaware" subgroups, the differences melted away.

Says LaBar, "The apparent deficit with aging was really just an artifact of a greater proportion of unaware participants in the older group's average." By comparing young apples to old apples and young oranges to old oranges, the Duke researchers were able to see that regardless of age, more aware participants became anxious at one rate--and less aware participants became anxious at another rate. It was not age, but rather awareness, that differentiated rates of learning.

Complex contingencies

Other researchers note differences between the study's findings and those in other, similar studies. For example, Temple University neuropsychologist Diana Woodruff-Pak, PhD, distinguishes between two different types of learned response: autonomic, or slow, responses such as skin conductance, versus somatic, or fast, responses such as the reflexive eyeblink response to a puff of air. Research into the conditioned eyeblink has found that conditioning slowed with age, and awareness didn't account for all the difference.

She points out that other eyeblink experiments have used a 400-millisecond CS, whereas the Duke team presented its CS for 10 times as long--four seconds. Woodruff-Pak suggests that the longer time interval gave participants more time to gain insight, thus elevating their rate of response.

LaBar wonders whether, for conditioning emotional reactions, "timing is not so critical because the autonomic nervous system is sluggish." It may just be apples and oranges; as researchers well know, the study of learning can be maddeningly complicated. Bellebaum and Daum conclude only that methodological and procedural issues make it hard to directly compare how fear conditioning differs from motor conditioning as we age.

In any case, the Duke researchers felt confident that once they took contingency awareness and nonassociative arousability (which is higher in the young) into account, fear conditioning is preserved in healthy aging. The finding meshes with other animal studies whose evidence is consistent with mechanisms for the conservation of fear learning across species as they age.

"Older adults can still make basic emotional associations to stimuli that are relevant to biological health," says LaBar. "Although they may not remember where they parked the car, they tend to remember that touching a hot stove is not a pleasant experience."

Neural stability, testing ability

Fear conditioning might hold up because the amygdala--a key brain area in processing perception and emotional behavior--doesn't shrink much with age, as do some other brain regions such as the cerebral cortex. In healthy aging, the amygdala (an ancient temporal-lobe structure below the cortex, next to the hippocampus) shrinks in volume by only 2 percent. The authors reflect that the "relative structural integrity of the aging amygdala appears to be sufficient to mediate simple forms of fear conditioning."

Understanding healthy aging may help to improve dementia tests, which currently use explicit learning (learning with awareness) to assess declarative, or factual, memory. LaBar and his colleagues, in studying nondeclarative, emotional memory, confirmed that awareness can confound results. In particular, conditioning tests that use two cues might be problematic because they tap both implicit and explicit learning; having both processes at work can muddy the diagnostic waters.

A more pure memory test should keep it simple. "If researchers want to test the impact of aging on nondeclarative memory, it may be better to use single-cue paradigms instead," asserts LaBar.

At the same time, he speculates that older adults might perform better on memory tests that emphasize events of a personal or biological relevance. After all, notes LaBar, "If they were unable to remember when they engaged in risky behavior or never learned how to avoid negative consequences, they probably would not have lived so long!"

Bellebaum and Daum speculate about broader implications of the Duke findings, given that eyeblink conditioning in older participants has helped to predict the later development of dementia. Whether fear conditioning has similar implications remains to be investigated, they note. Beyond testing, LaBar says that if the findings are extended to positive conditioning, they may help people with declarative memory loss. "For example, to train a forgetful elderly person to put her keys in a certain spot, you might succeed by making it more personally relevant--maybe by placing the keys near a picture of a beloved grandchild," he explains.

"Thus," he continues, "we could capitalize on the emotional fitness of the elderly to help them circumvent other age-associated declines in cognition and behavior."

Rachel Adelson is a science writer in Raleigh, N.C.