In the past few months, several studies have come out showing the value of three different tests of people's sense of smell for improving the accuracy of MCI and Alzheimer's diagnosis, or pointing to increased risk. The studies also add to growing evidence that a decline in sense of smell is an early marker for mild cognitive impairment and Alzheimer’s. Indeed, it appears that this sensory loss is a very early symptom, preceding even the shrinking of the entorhinal cortex (the first brain region to show signs of atrophy).

Smell test improves accuracy of MCI & Alzheimer's diagnosis

A simple, commercially available test known as the Sniffin' Sticks Odor Identification Test, in which subjects must try to identify 16 different odors, was given to 728 older adults, as well as a standard cognitive test (the Montreal Cognitive Assessment).

The participants had already been evaluated by doctors and classified as being healthy (292 subjects), having MCI (174: 150 aMCI, 24 naMCI), or having Alzheimer's (262).

It was found that, while the cognitive test alone correctly classified 75% of people with MCI, the number rose to 87% when the sniff test results were added. Diagnosis of Alzheimer's, and of subtypes within MCI, was also improved.

The smell test normally takes 5 to 8 minutes to administer; the researchers are trying to get it down to 3 minutes, to encourage greater use.

A new smell test

Another recent study validates a new smell test which is rather more complicated. The test was developed because the standard University of Pennsylvania Smell Identification Test doesn’t take into account the great variation in olfactory ability among healthy individuals. The ability of normal individuals to recognize and discriminate between odors can vary by as much as 40 times!

The new test is actually four tests:

  • In the OPID (Odor Percept IDentification)-10 test, participants are presented with 10 odors (menthol, clove, leather, strawberry, lilac, pineapple, smoke, soap, grape, lemon) for two seconds each. They are then asked whether the scent is familiar and given a choice of four of the 10 words from which are asked to pick the best one that describes the odor.
  • The Odor Awareness Scale (OAS) assesses their overall attention to environmental odors and how they are affected emotionally and behaviorally by scents.
  • The OPID-20 test includes an additional 10 odors (banana, garlic, cherry, baby powder, grass, fruit punch, peach, chocolate, dirt, orange). Participants are first asked whether a presented odor was included in the OPID-10 test and then asked which word best describes the odor. Their ability to remember odors from the first test determines their POEM (Percepts of Odor Episodic Memory) score.
  • In the Odor Discrimination (OD) test, participants are presented with two consecutive odors and asked whether they were different or the same, a process that is repeated 12 times with different paired scents.

The study involved 183 older adults, of whom 70 were cognitively normal, 74 tested normal but were concerned about their cognitive abilities, 29 had MCI and 10 had been diagnosed with possible or probable Alzheimer's disease.

Results of the OPID-20 test significantly differentiated among the four groups of participants, and those results correlated with the thinning of the hippocampus and the entorhinal cortex. Participants' ability to remember a previously presented aroma, as reflected in the POEM score, was also significant, with participants with Alzheimer's disease performing at no better than chance.

POEM scores of the two cognitively normal groups were compared with what would have been predicted based on their ability to identify and differentiate between odors, as reflected in the OAS and OD tests. Poor POEM performers were more likely to have the ‘Alzheimer's gene’ (APOEe4), showed thinning of the entorhinal cortex, and poorer cognitive performance over time.

Validation of UPSIT

However, two 2016 studies support the use of the University of Pennsylvania Smell Identification Test (UPSIT), and suggest it may offer a practical, low-cost alternative to other tests.

In one study, UPSIT was administered to 397 older adults (average age 80) without dementia, who were also given an MRI scan to measure the thickness of the entorhinal cortex (the first brain region to be affected by Alzheimer's disease). After four years, 50 participants (12.6%) had developed dementia, and nearly 20% had signs of cognitive decline.

Low UPSIT scores, but not entorhinal cortical thickness, were significantly associated with dementia and Alzheimer's disease, and with cognitive impairment. Entorhinal cortical thickness was significantly associated with UPSIT score in those who transitioned from MCI to dementia.

In other words, it looks like impairment in odor identification precedes thinning in the entorhinal cortex.

In another study, UPSITwas administered to 84 older adults, of whom 58 had MCI, as well as either beta amyloid PET scanning or analysis of cerebrospinal fluid. After six months, 67% had signs of memory decline, and this was predicted by amyloid-beta levels (assessed by either method), but not UPSIT score. However, participants with a score of less than 35 were more than three times as likely to have memory decline as those with higher UPSIT scores.

The researchers suggest the association wasn’t as strong in this study because of the younger age of participants (median age 71), their higher education, and the short follow-up.

[4209] Quarmley M, Moberg PJ, Mechanic-Hamilton D, Kabadi S, Arnold SE, Wolk DA, Roalf DR. Odor Identification Screening Improves Diagnostic Classification in Incipient Alzheimer’s Disease. Journal of Alzheimer's Disease [Internet]. 2017 ;55(4):1497 - 1507. Available from:

[4210] Dhilla AAlefiya, Asafu-Adjei J, Delaney MK, Kelly KE, Gomez-Isla T, Blacker D, Johnson KA, Sperling RA, Hyman BT, Betensky RA, et al. Episodic memory of odors stratifies Alzheimer biomarkers in normal elderly. Annals of Neurology [Internet]. 2016 ;80(6):846 - 857. Available from:

Lee, Seonjoo et al. 2016. Predictive Utility of Entorhinal Cortex Thinning and Odor Identification Test for Transition to Dementia and Cognitive Decline in an Urban Community Population. Presented at the Alzheimer's Association's International Conference in Toronto.

Kreisl, William et al. 2016. Both Odor Identification and Amyloid Status Predict Memory Decline in Older Adults. Presented at the Alzheimer's Association's International Conference in Toronto.

A couple of studies reported at the recent Annual Conference of the British Psychological Society are intriguing.

In the first study, 180 healthy adults completed questionnaires relating to their mood before being given either a drink of peppermint tea, chamomile tea or hot water. Twenty minutes later, their memory and cognition were tested, followed by another mood questionnaire.

Peppermint tea significantly improved long term memory, working memory and alertness compared to both chamomile and hot water, while chamomile tea (consistent with its reputed calming/sedative effect) significantly slowed memory and attention speed compared to both peppermint and hot water.

In the second study, 150 older adults (65+) were tested on their prospective memory while in one of three rooms: a room that was scented either with rosemary or lavender, or an unscented room.

The scented rooms had four drops of either lavender or peppermint essential oils placed on an aroma stream fan diffuser, switched on five minutes before the participants entered the room. Prospective memory was tested by needing to remember to pass on a message at a given time during the procedure, and having to switch tasks when a specific event occurred.

The room scented with rosemary significantly enhanced prospective memory compared to the room with no aroma. It also significantly increased alertness, while lavender significantly increased calmness and contentedness.

As I said, these are conference papers, and I know no more than revealed in the press release. However, these strategies are easy and harmless enough that you might want to try them for yourself.

Bussey, L. 2016. I really must post that letter! Aromas of essential oils impact on prospective memory in an older cohort. Presented at the British Psychological Society's 2016 Annual Conference in Nottingham.

Moss, M. 2016. Contrasting Effects of Peppermint and Chamomile Tea on Cognition and Mood. Presented at the British Psychological Society's 2016 Annual Conference in Nottingham.

A small study involving 20 people has found that those who were exposed to 1,8-cineole, one of the main chemical components of rosemary essential oil, performed better on mental arithmetic tasks. Moreover, there was a dose-dependent relationship — higher blood concentrations of the chemical were associated with greater speed and accuracy.

Participants were given two types of test: serial subtraction and rapid visual information processing. These tests took place in a cubicle smelling of rosemary. Participants sat in the cubicle for either 4, 6, 8, or 10 minutes before taking the tests (this was in order to get a range of blood concentrations). Mood was assessed both before and after, and blood was tested at the end of the session.

While blood levels of the chemical correlated with accuracy and speed on both tasks, the effects were significant only for the mental arithmetic task.

Participants didn’t know that the scent was part of the study, and those who asked about it were told it was left over from a previous study.

There was no clear evidence that the chemical improved attention, but there was a significant association with one aspect of mood, with higher levels of the scent correlating with greater contentment. Contentment was the only aspect of mood that showed such a link.

It’s suggested that this chemical compound may affect learning through its inhibiting effect on acetylcholinesterase (an important enzyme in the development of Alzheimer's disease). Most Alzheimer’s drugs are cholinesterase inhibitors.

While this is very interesting (although obviously a larger study needs to confirm the findings), what I would like to see is the effects on more prolonged mental efforts. It’s also a little baffling to find the effect being limited to only one of these tasks, given that both involve attention and working memory. I would also like to see the rosemary-infused cubicle compared to some other pleasant smell.

Interestingly, a very recent study also suggests the importance of individual differences. A rat study compared the effects of amphetamines and caffeine on cognitive effort. First of all, giving the rats the choice of easy or hard visuospatial discriminations revealed that, as with humans, individuals could be divided into those who tended to choose difficult trials (“workers”) and those who preferred easy ones (“slackers”). (Easy trials took less effort, but earned commensurately smaller reward.)

Amphetamine, it was found, made the slackers worked harder, but made the workers take it easier. Caffeine, too, made the workers slack off, but had no effect on slackers.

The extent to which this applies to humans is of course unknown, but the idea that your attitude to cognitive effort might change how stimulants affect you is an intriguing one. And of course this is a more general reminder that factors, whatever they are, have varying effects on individuals. This is why it’s so important to have a large sample size, and why, as an individual, you can’t automatically assume that something will benefit you, whatever the research says.

But in the case of rosemary oil, I can’t see any downside! Try it out; maybe it will help.

The olfactory bulb is in the oldest part of our brain. It connects directly to the amygdala (our ‘emotion center’) and our prefrontal cortex, giving smells a more direct pathway to memory than our other senses. But the olfactory bulb is only part of the system processing smells. It projects to several other regions, all of which are together called the primary olfactory cortex, and of which the most prominent member is the piriform cortex. More recently, however, it has been suggested that it would be more useful to regard the olfactory bulb as the primary olfactory cortex (primary in the sense that it is first), while the piriform cortex should be regarded as association cortex — meaning that it integrates sensory information with ‘higher-order’ (cognitive, contextual, and behavioral) information.

Testing this hypothesis, a new rat study has found that, when rats were given training to distinguish various odors, each smell produced a different pattern of electrical activity in the olfactory bulb. However, only those smells that the rat could distinguish from others were reflected in distinct patterns of brain activity in the anterior piriform cortex, while smells that the rat couldn’t differentiate produced identical brain activity patterns there. Interestingly, the smells that the rats could easily distinguish were ones in which one of the ten components in the target odor had been replaced with a new component. The smells they found difficult to distinguish were those in which a component had simply been deleted.

When a new group of rats was given additional training (8 days vs the 2 days given the original group), they eventually learned to discriminate between the odors the first animals couldn’t distinguish, and this was reflected in distinct patterns of brain activity in the anterior piriform cortex. When a third group were taught to ignore the difference between odors the first rats could readily distinguish, they became unable to tell the odors apart, and similar patterns of brain activity were produced in the piriform cortex.

The effects of training were also quite stable — they were still evident after two weeks.

These findings support the idea of the piriform cortex as association cortex. It is here that experience modified neuronal activity. In the olfactory bulb, where all the various odors were reflected in different patterns of activity right from the beginning (meaning that this part of the brain could discriminate between odors that the rat itself couldn’t distinguish), training made no difference to the patterns of activity.

Having said that, it should be noted that this is not entirely consistent with previous research. Several studies have found that odor training produces changes in the representations in the olfactory bulb. The difference may lie in the method of neural recording.

How far does this generalize to the human brain? Human studies have suggested that odors are represented in the posterior piriform cortex rather than the anterior piriform cortex. They have also suggested that the anterior piriform cortex is involved in expectations relating to the smells, rather than representing the smells themselves. Whether these differences reflect species differences, task differences, or methodological differences, remains to be seen.

But whether or not the same exact regions are involved, there are practical implications we can consider. The findings do suggest that one road to olfactory impairment is through neglect — if you learn to ignore differences between smells, you will become increasingly less able to do so. An impaired sense of smell has been found in Alzheimer’s disease, Parkinson's disease, schizophrenia, and even normal aging. While some of that may well reflect impairment earlier in the perception process, some of it may reflect the consequences of neglect. The burning question is, then, would it be possible to restore smell function through odor training?

I’d really like to see this study replicated with old rats.

190-million-year-old fossil skulls of Morganucodon and Hadrocodium, two of the earliest known mammal species, has revealed that even at this early stage of mammalian evolution, mammals had larger brains than would be expected for their body size. High-resolution CT scans of the skulls have now shown that this increase in brain size can be attributed to an increase in those regions dealing with smell and touch (mammals have a uniquely well developed ability to sense touch through their fur).

Comparison of these fossils with seven fossils of early reptiles (close relatives of the first mammals), 27 other primitive mammals, and 270 living mammals, has further revealed that the size of the mammalian brain evolved in three major stages. First, an initial increase in the olfactory bulb and related areas (including the cerebellum) by 190 million years ago; then another jump in the size of these regions shortly after that time; and finally an increase in those regions that control neuromuscular coordination by integrating different senses by 65 million years ago.

It’s speculated that the initial increase in smell and touch was driven by early mammals being nocturnal — dinosaurs being active during the day.

Previous research suggesting loss of smell function may serve as an early marker of Alzheimer's disease has now been supported by a finding that in genetically engineered mice, loss of smell function is associated with amyloid-beta accumulation in the brain, and that amyloid pathology occurs first in the olfactory region. It was striking how sensitive olfactory performance was to even the smallest amount of amyloid presence in the brain as early as three months of age (equivalent to a young adult).

Previous research suggesting loss of smell function may serve as an early marker of Alzheimer's disease has now been supported by a finding that in genetically engineered mice, loss of smell function is associated with amyloid-beta accumulation in the brain, and that amyloid pathology occurs first in the olfactory region. It was striking how sensitive olfactory performance was to even the smallest amount of amyloid presence in the brain as early as three months of age (equivalent to a young adult).

The effect of smell on learning and memory was investigated in an experiment that used three different ambient odors (osmanthus, peppermint, and pine).

Osmanthus was used to see whether there was a difference in performance depending on whether the smell was novel or familiar. Peppermint and pine were used to see whether the appropriateness or inappropriateness of the smell made a difference to memory.

In the experiment, subjects were individually shown into a room in which the odor was present. Their attention was called to the smell, and to ensure their attention to the smell, they were given a questionnaire to fill out about the room environment. They were left alone in the room for ten minutes to promote encoding of contextual cues.

The experimenter then read out a list of 20 common nouns, pausing after each one for the subject to describe an event that the word reminded them of. Memory for the words was tested 48 hours later.

It was found that word recall was best when the novel odor (osmanthus) was present during learning and again at testing. Among the familiar odors, recall was better if the smell was contextually inappropriate (peppermint). The improvement in recall only occurs when the odor is present at both encoding (learning) and retrieval (testing). Clearly, smell is a good contextual cue.

Herz, R.S. (1997). The effects of cue distinctiveness on odor-based context dependent memory. Memory and Cognition, 25, 375-380.

Older news items (pre-2010) brought over from the old website

Why smells can be so memorable

Confirming the common experience of the strength with which certain smells can evoke emotions or memories, an imaging study has found that, when people were presented with a visual object together with one, and later with a second, set of pleasant and unpleasant odors and sounds, then presented with the same objects a week later, there was unique activation in particular brain regions in the case of their first olfactory (but not auditory) associations. This unique signature existed in the hippocampus regardless of how strong the memory was — that is, it was specific to olfactory associations. Regardless of whether they were smelled or heard, people remembered early associations more clearly when they were unpleasant.

[2543] Yeshurun Y, Lapid H, Dudai Y, Sobel N. The Privileged Brain Representation of First Olfactory Associations. Current Biology [Internet]. 2009 ;19:1869 - 1874. Available from:

Difficulty identifying odors may predict cognitive decline

Older adults who have difficulty identifying common odors may have a greater risk of developing mild cognitive impairment, increasingly recognized as a precursor to Alzheimer’s disease.  A study of nearly 600 older adults (average age 79.9) found that 30.1% developed mild cognitive impairment over the five-year period of the study. Risk of developing mild cognitive impairment was greater for those who scored worse on an odor identification test given at the start of the study. For example, those who scored below average (eight) were 50% more likely to develop MCI than those who scored above average (11). This association did not change when stroke, smoking habits or other factors that might influence smell or cognitive ability were considered. Impaired odor identification was also associated with lower cognitive scores at the beginning of the study and with a more rapid decline in episodic memory (memory of past experiences), semantic memory (memory of words and symbols) and perceptual speed. The odor test involved identifying 12 familiar odors given four possible alternatives to choose from.

[1130] Wilson RS, Schneider JA, Arnold SE, Tang Y, Boyle PA, Bennett DA. Olfactory Identification and Incidence of Mild Cognitive Impairment in Older Age. Arch Gen Psychiatry [Internet]. 2007 ;64(7):802 - 808. Available from:

Odor can help memory, in some circumstances

A study in which students played a computer version of a common memory game in which you turn over pairs of cards to find each one's match found that those who played in a rose-scented room and were later exposed to the same scent during slow-wave sleep, remembered the locations of the cards significantly better than people who didn't have that experience (97% vs 86%). Those exposed to the odor during REM sleep, however, saw no memory boost. Imaging revealed the hippocampus was activated when the odor was presented during slow-wave sleep. Having the smell available throughout sleep wouldn’t help, however, because we adapt to smells very quickly. Being exposed to the smell when being tested didn’t help either. Nor did experiencing the odor during slow-wave sleep help when the memory task involved a different type of memory — learning a finger-tapping sequence — probably because procedural memory doesn’t depend on the hippocampus.

[1206] Rasch B, Buchel C, Gais S, Born J. Odor Cues During Slow-Wave Sleep Prompt Declarative Memory Consolidation. Science [Internet]. 2007 ;315(5817):1426 - 1429. Available from:

Scent of fear impacts cognitive performance

A study involving 75 female students found that those who were exposed to chemicals from fear-induced sweat performed more accurately on word-association tasks than did women exposed to chemicals from other types of sweat or no sweat at all. When processing meaningfully related word pairs, the participants exposed to the fear chemicals were significantly more accurate than those in either the neutral sweat or the control (no-sweat) condition. When processing word pairs that were ambiguous in threat content, such as one neutral word paired with a threatening word or a pair of neutral words, subjects in the fear condition were significantly slower in responding than those in the neutral sweat condition.

Chen, D., Katdare, A. & Lucas, N. 2006. Chemosignals of Fear Enhance Cognitive Performance in Humans. Chemical Senses, Advance Access published on March 9, 2006

Brain region involved in recalling memories from smell identified

We all know the power of smell in triggering the recall of memories. New research has found the specific area of the brain involved in this process - a section of the hippocampus called CA3. The hippocampus has long been known to play a crucial part in forming new memories. It appears that the CA3 region of the hippocampus is crucial for recalling memories from partial representations of the original stimulus.

[1060] Wilson MA, Tonegawa S, Nakazawa K, Quirk MC, Chitwood RA, Watanabe M, Yeckel MF, Sun LD, Kato A, Carr CA, et al. Requirement for Hippocampal CA3 NMDA Receptors in Associative Memory Recall. Science [Internet]. 2002 ;297(5579):211 - 218. Available from: