Hedonic Assessment of Odors: A Comparison of Two Sensory Scales for Use with Alzheimer’s Disease Patients and Elderly Individuals

Abstract

Background:

Several clinical studies concerning the olfactory function of patients with cognitive impairment have used sensory scales to investigate hedonic perception. However, no study has focused on the choice of the most appropriate sensory hedonic scale for the individuals with neurodegenerative disorders or other psychiatric diseases involving cognitive deficits.

Objective:

The aim of this study was to investigate the ability of patients with Alzheimer’s disease (AD) to use two hedonic scales (category scale and linear scale) and compare their discriminatory capacity, repeatability, and ease of use. This should allow us to identify the most appropriate hedonic scale for patients with AD.

Methods:

We recruited 18 patients with mild to moderate AD, and 20 healthy volunteers matched for gender, age, smoking status, and educational level. The participants underwent a clinical assessment and hedonic evaluation of three odorants (pleasant, unpleasant, and neutral), using a five-point category scale and a 10-cm linear scale with a marked mid-point.

Results:

AD patients were able to use hedonic scales as well as paired healthy elderly subjects. The linear scale performed slightly better in terms of ease of use for both patients and healthy controls and discriminatory capacity for AD patients. The results for AD patients and controls with both scales were repeatable.

Conclusion:

The linear scale may be more appropriate for AD patients pending further studies involving a larger population of patients, using several odorants.

Keywords

Alzheimer’s disease five-point category scale hedonic evaluation linear scale olfaction

INTRODUCTION

Olfactory dysfunction is one of the earliest signs of Alzheimer’s disease (AD) [1, 2]. The deficit of several olfactory functions in this disease is due to the overlapping of brain structures involved in olfactory processing and those altered in AD: olfactory epithelium, olfactory bulb, hippocampus, amygdala, piriform cortex, anterior cingular cortex, orbito frontal cortex, and habenula [3, 4]. Several validated and standardized psychophysical tests have been used to evaluate olfactory deficits in patients with cognitive disorders: they measure various olfactory-related components, such as odor detection threshold, identification, discrimination, and recognition memory [5, 6]. Olfactory deficits have been reported in AD patients for these olfactory-related components (for reviews see [3, 7]).

Scales for the evaluation of various odorant characteristics, such as pleasantness, intensity, and familiarity, can be used in the clinical field to study olfactory function [8 –10]. Among these three olfactory parameters, the hedonic aspect (pleasantness) has a central place. Indeed, hedonic pleasure contributes to improving the affective state, social relations, and the enjoyment of life (through the perception of pleasant odors in food or the environment). Furthermore, a pertinent hedonic evaluation could be used to improve food consumption in the elderly and/or in patients with neurodegenerative diseases to guide recommendations for more suitable food. The hedonic valence of an odor can influence the subject’s ability to evaluate various olfactory functions, such as odor identification, intensity perception, and discrimination [11, 12]. Thus, the study of pleasantness in AD is essential, but it has only been poorly investigated in AD patients (only three studies). Indeed, Joussain et al. [10] demonstrated significantly lower hedonic ratings for AD patients than controls for pleasant odorants, whereas there was no difference for neutral or unpleasant odorants. The two other studies did not observe the difference between AD patients and controls concerning hedonic responses [13, 14]. The small number of inconsistent studies concerning the investigation of hedonic perception of odorants in AD may be due to the lack of a suitable and/or standardized scale for the clinical population with cognitive deficits.

Several scaling methods and scales have been developed in the psychophysics field to evaluate hedonic perception. There are excellent reviews concerning the theoretical and practical differences among methods and their potential advantages and disadvantages (for reviews, see [15, 16]). We will focus on scaling methods used in AD and psychiatric diseases involving cognitive deficits and some of their advantages and disadvantages.

Two types of scales have been used to study the perception of an odors’ pleasantness in AD patients: 1) a 10-cm linear hedonic scale labeled at each end (highly unpleasant/very pleasant) [13, 14], and 2) a rating from 1 (very unpleasant) to 9 (very pleasant); i.e., the subjects were asked to attribute the notes [10]. The linear hedonic scale has the advantage of giving subjects more freedom to express their sensory perception and accurately assess the pleasantness, because it is possible to mark any point on the scale [16, 17]. Besides, compared to the category scales, the linear scale is less used because it is more difficult to understand by subjects, slowing down the sensory evaluation [16, 18]. The rating method from 1 to 9 without using a visual scale has been little used for hedonic evaluation. There have been no studies comparing this method with other frequently used scaling methods. Other scales used in psychiatric diseases involving cognitive deficits (major depression, bipolar disorder, and schizophrenia) are five- or seven-point hedonic scales based on categories, which range from dislike a lot to like a lot [19 –21] and the linear regularly segmented and numbered rating scale [22]. This last scale is similar to the hybrid hedonic scale (presented below) and results from the combination of structured and unstructured linear scales. Category scales are the most widely used in sensory analysis for acceptance testing (hedonic evaluation). They are easily understood by the subjects and require no prior training [15, 16]. These are scales in which the degree of acceptance (pleasantness) is evaluated through verbal and/or numerical labels or “faces” (facial hedonic scale). There are also three-, five-, seven-, and nine-point hedonic scales [23]. The nine-point hedonic scale is the most used in consumer studies. Scales with fewer categories (five- or seven-point hedonic scales) have been used in studies carried out with clinical populations or with five- to eight-year-old children, mainly for individuals with limited cognitive abilities [23]. Finally, the hybrid hedonic scale is anchored with verbal affective labels in the middle and extreme regions of the scale, making it easier to use by the subjects than the unstructured scale. The remaining portions are marked with equidistant points to better define the degree and direction of hedonic continuity. This scale provides better discriminatory power since it is not restricted to a limited number of categories [16, 17]. In contrast to the linear scale and the category scale, the hybrid hedonic scale has only been recently developed and there are not yet sufficient available data to confirm its advantages.

All scaling methods used in AD and psychiatric diseases mentioned above were arbitrarily chosen without any justification based on literature review or pre-tests carried out to choose the most appropriate scale for the patients participating in the studies. However, the psychophysical literature of the olfactory/gustatory field provides clear evidence that the choice of scaling technique can considerably influence the reliability of the results, depending on the experimental conditions (for example, the experimental context, the stimulus range, or the type of subject: children, young adults, or elderly) [15 , 24–28]. Indeed, no study has investigated the advantages and drawbacks of the sensory scales for olfactory/gustatory hedonic evaluation for patients with AD.

The general aim of the present study was to gain a better understanding of hedonic olfactory perception in AD based on the various aspects discussed above. This involved the investigation of two objectives. The first was to study the ability of AD patients to use two hedonic scales already used for clinical populations with cognitive deficits: the five-point category scale and the linear scale. The second was to determine which of these two sensory scales is more appropriate for AD patients. The two scales were compared on the basis of three criteria: 1) repeatability of the subjects’ hedonic evaluations, 2) discriminatory capacity concerning the three odorants of different hedonic valence, and 3) ease of use.

MATERIALS AND METHODS

Participants

Eighteen subjects with mild to moderate AD (mean age: 81.5±5.9) according to the criteria of McKhann et al. [29], including clinical findings, neuropsychological evaluations, and brain imaging were recruited at the university hospital of Tours, France (Centre Mémoire de Ressources et de Recherche, CMRR). Patients were required to score no lower than 15/30 on the Mini-Mental State Examination (MMSE) [30] (mean MMSE score: 19.4±3.2; range: 15–25) for inclusion in the study. This threshold was chosen as patients with this score can participate in the tests and give informed consent.

A control sample of 20 healthy volunteers (mean age: 80.9±5.9) with no history of mental or neurodegenerative illness was included. These subjects were matched with the patients for gender (χ² = 0.04; df = 1; p = 0.8), age (t = –0.5; df = 36; p = 0.6), educational level (t = 0.1; df = 36; p = 0.9), and smoking status. Their average MMSE score was 28.4±0.9 (range: 27–30).

Exclusion criteria for all subjects included head injury, current substance abuse, odor allergy, a current cold, or any alteration to their sense of smell due to nasal cavity inflammation, nasal polyposis, nasal surgery, etc. The inclusion criteria included the absence of anosmia to the used odorants. Indeed, all subjects who could not perceive one or several odors during the olfactory test were excluded from the study (this was the case for one healthy subject who had difficulties perceiving one of the three odorantsused).

The characteristics of the two groups are presented in Table 1.

Table 1

Characteristics of the two groups of subjects

	AD patients (n = 18)	Healthy controls (n = 20)
Number of women/men	13/5	15/5
Mean age, years (SD)	81.5 (5.9)	80.9 (5.9)
Age range, years	71–93	73–97
Non-smokers/smokers	16/2	18/2
MMSE, mean score (SD)	19.4 (3.2)	28.4 (0.9)
MMSE, score range	15–25	27–30
Socio-educational level, mean score^*(SD)	1.3 (0.6)	1.4 (0.6)

^*Socio-educational level was calculated on a three-point scale (1, 2, and 3, corresponding to primary, middle, and high school levels, respectively). MMSE, Mini-Mental State Examination.

Olfactory stimuli

Three odors with different hedonic valences were chosen, because discriminatory capacity was one of the three criteria used to assess the efficacy of the scaling methods: benzaldehyde (BEN) with a pleasant odor of bitter almond (0.5 ml/l), 1-octen-3-ol (OCT) with a neutral odor of mushrooms (0.15 ml/l), and butyric acid (BUT) with an unpleasant rancid odor (1.6 ml/l) [22 , 31–33]. All odorants were supplied by Fisher Scientific (France). Their supra-threshold concentrations were chosen to be approximately iso-intense. The odorants were diluted in mineral oil (Sigma-Aldrich, France). The odorant solutions were poured into 60 ml brown glass flasks (10 ml per flask), and each flask coded with a random three-digit number.

Scaling methods

Two scaling methods of hedonic assessment were used: (i) a five-point category scale: 1 = very unpleasant, 2 = unpleasant, 3 = neither pleasant nor unpleasant, 4 = pleasant, and 5 = very pleasant and (ii) a 10-cm linear scale anchored on the left by 0, very unpleasant; in the middle by 5, neither pleasant nor unpleasant, and on the right by 10, very pleasant (Fig. 1). When being evaluated by the five-point category scale, the subject was asked to smell each odor and evaluate its degree of pleasantness by placing a cross in the box corresponding to her/his perception. For the linear scale, the subject was instructed to indicate the pleasantness of the odor by placing a vertical line on the scale.

Fig.1

Sensory scales used for hedonic ratings (A: linear scale, B: five-point category scale).

These two scales were selected after a pre-test carried out on eight patients with mild to moderate AD. Four scales (a linear scale, hybrid hedonic scale, and five- and seven-point hedonic scales) were presented to the subjects and they were asked to rank them according to ease of use or understandability (rank 1: easier to use; rank 4: less easy to use). The obtained sums of ranks for each scale were: linear scale = 28; five-point scale = 26; hybrid scale = 16 and seven-point scale = 10. The Freidman rank test revealed a significant difference between the scales (Q = 16.2; df = 3; p = 0.001). The post-hoc Nemenyi procedure showed that subjects considered the linear scale, five-point hedonic scale, and hybrid scale to be easier to use than the seven-point hedonic scale. However, the hybrid hedonic scale was judged not to be significantly different from the three other scales. Besides, six of the eight AD patients have preconized to mark the mid-point of the linear scale. Therefore, the five-point hedonic scale and the linear scale with a marked mid-point were chosen for the present study.

Experimental procedure

The study was approved by the local ethics committee (CPP Tours Ouest-1, France) and was conducted in accordance with good clinical practice and the Declaration of Helsinki.

Upon recruitment, all subjects were informed that they would have to smell various odors that can be used to flavor food and/or cosmetic products. Prior to the measurement session, the sensory task (evaluation of odor pleasantness) and the two scaling methods were explained to the participants and a brief demonstration was given, if necessary. In addition to the oral information, written instructions on how to perform the tests was available in the questionnaire.

Measurements were performed during a single session of 20 min for each participant. The tests were performed individually. A monadic sequential stimuli presentation was used for the two methods. The presentation order of the stimuli was randomized and was identical for patients and controls and the two methods to facilitate comparison. The mushroom odor (OCT) was duplicated, and two flasks presented to study repeatability. Only one neutral odor was used to test this aspect to limit olfactory saturation and fatigue of the AD patients and avoid the perseverance effect. The order of presentation of the two scaling methods was random and was identical for patients and controls. The subjects were helped during the sessions if they had questions.

Every individual optimizes his or her sniffing technique to obtain his or her maximum sensitivity [34]. Therefore, the subjects were not limited in the time allowed for sniffing, but a one-minute interval between samples was imposed to prevent olfactory adaptation and saturation.

After completion of the sensory task, the subjects were asked which of the scales they found easier to use. The score of each group for this task was recorded as the number of responses, which was one of the three criteria used to assess the efficacy of the two scaling methods. All participants were also asked to orally explain their choice.

Statistical analysis

The responses obtained using the linear scale were recorded as a score from 0 to 10. The hedonic scores obtained with the category scale were recorded as values from 0 to 4. The category values for the five-point hedonic scale were multiplied by 2.5 to give a range from 0 to 10 to enable direct comparisons. Mean scores and standard deviations for hedonic responses were calculated for patient and control groupsseparately.

Two-tailed unpaired Student’s t-tests were used to compare age and educational level between the groups. The chi-square test was used to compare the gender distribution between the groups.

The two scales were compared concerning the repeatability of the subjects’ hedonic evaluations, the ability of subjects to discriminate between the three odorants of different hedonic valence, and the ease of use for the two groups of subjects.

Repeatability of hedonic evaluation

Two-tailed paired Student’s tests were used to study the test-retest reliability of the hedonic responses recorded by the two methods.

Hedonic evaluation and discrimination of odors

Two-way analysis of variance (ANOVA) was carried out to determine the overall significance for the main effects (group: patients and controls; odor: BEN, OCT, and BUT), and their interaction (group*odor) for each scale separately. Odor discriminatory capacity was investigated by exploring the number of stimuli that each group identified to be different from one another: ANOVA with Bonferroni multiple t-tests were used to compare all possible pairs of odors for each group, adjusting the results to reflect the multiple tests of significance (Bonferroni correction: the level of significance was set at p = 0.003). This last analysis also allowed comparison of the hedonic evaluation of each odor between the two groups of subjects.

Ease of use of the scale

The z test was used for each group to compare the proportions of subjects preferring the category scale or the linear scale according to their ease of use.

The correlation of the individual hedonic responses to odors as recorded by the two methods was examined using Pearson’s correlation coefficient.

All statistical analyses were performed using XLSTATâ-Pro, release 5.2 at ^®= 5%.

RESULTS

Repeatability of hedonic evaluation

We analyzed the test-retest reliability of the subjects’ responses. Both groups responded consistently to the repeated OCT stimulus using both the linear scale (two-tailed paired Student’s test; AD patients: t = –0.9; df = 17; p = 0.4; Controls: t = –1.3; df = 19; p = 0.21) and five-point category scale (two-tailed paired Student’s test; AD patients: t = 0.3; df = 17; p = 0.77; Controls: t = –1.5; df = 19; p = 0.16) (Table 2).

Table 2

Mean (SD) values of hedonic responses for AD patients and healthy controls. R1: first repetition; R2: second repetition; (p-value: two-tailed paired Student test)

	Linear scale			Category scale
	R1	R2	p-value	R1	R2	p-value
AD patients	4.9 (1.2)	5.4 (2.1)	0.40	5.0 (1.9)	4.9 (1.8)	0.77
Healthy controls	3.4 (1.8)	3.8 (1.6)	0.21	3.5 (2.1)	4.3 (2.0)	0.16

Hedonic evaluation and discrimination of odors

For the linear scale, there was a significant effect for each group and odor (F _(1,36) = 13.6, p < 0.001; F _(2,639) = 121.8, p < 0.001, respectively). There was no significant group*odor interaction (F _(2,5) = 0.9, p = 0.39). The Bonferroni t-tests showed that both groups significantly differentiated the three odors according to their hedonic quality (Fig. 2). For the category scale, there was a significant effect of each group and odor, and there was a significant interaction (F _(1,14) = 4.2, p = 0.04; F _(2,385) = 55.3, p < 0.001; F _(2,74) = 10.7, p < 0.001, respectively). Two-by-two multiple comparison of the means showed that the control subjects were able to discriminate the odors according to their hedonic valence. However, the AD patients evaluated the neutral odor (OCT) to not be significantly different from both pleasant and unpleasant odors. They perceived a significantdifference between BENZ and BUT only (Fig. 2).

Fig.2

Mean hedonic scores for the three odors evaluated by AD patients and healthy controls using the linear and five-point category scales. For each group of subjects and for each scale, means with the same letters were not significantly different (Bonferroni t-test). Vertical bars indicate the standard deviation of the mean.

There was no significant difference between the two groups of subjects concerning the hedonic evaluation of each odorant using the linear scale (Bonferroni multiple t-test). There was a significant difference for the unpleasant odorant (BUT) using the category scale, which was evaluated to be significantly less unpleasant by AD patients. There was no between-group difference for pleasant (BENZ) and neutral (OCT) odorants (Table 3).

Table 3

Comparison of the mean (SD) values of hedonic responses of AD patients and healthy controls for each odorant (p-value: Bonferroni multiple t-test)

Odorant	Linear scale			Category scale
	AD patients	Healthy controls	p-value	AD patients	Healthy controls	p-value
BENZ	7.8 (1.6)	7.2 (2.0)	0.3	5.8 (1.7)	7.4 (2.2)	0.012
OCT	4.9 (1.2)	3.4 (1.8)	0.004	5.0 (1.9)	3.5 (2.1)	0.015
BUT	2.3 (1.8)	1.1 (1.0)	0.02	3.2 (1.9)	1.0 (1.3)	<0.001

Ease of use of the scales

Control subjects found the linear scale easier to use (z = 4.7, p < 0.001). There was no significant difference between the ease of use of the two scales reported by the AD patients (z = –0.7, p = 0.50) (Fig. 3).

Fig.3

The proportions of subjects preferring the five-point category or the linear scale according to their ease to use, for AD patients (n = 18) and healthy controls (n = 20) (z test; ns: not significant; ^***p < 0.001).

Correlation between the two methods

The correlation between individual hedonic responses to odors recorded through use of the linear scale and category scale was highly significant for both AD patients (r = 0.59, p < 0.001;54 observations) and healthy controls (r = 0.85, p < 0.001; 60 observations).

DISCUSSION

The aim of the present study was twofold. First, we investigated the olfactory hedonic perception of AD patients by studying their ability to use a five-point category hedonic scale and a linear hedonic scale. Second, we compared these two sensory scales according to their discriminatory capacity, repeatability, and ease of use to identify which scale is more effective for patients with AD. We addressed these two objectives by studying the hedonic perception of three odorants of different hedonic valence and comparing the results of patients with mild to moderate AD with those of healthy subjects.

There was no difference in hedonic perception of pleasant, neutral, and unpleasant odors between the two groups using the linear scale. The hedonic perception of pleasant and neutral odors between AD patients and controls was also no different by the category scale, whereas the unpleasant odor was evaluated to be less unpleasant by the AD patients than the controls. These observations show that the type of scale used by AD patients can influence the results. Indeed, our results with the linear scale are in accordance with previous studies using a similar hedonic scale for AD patients with similar global cognitive efficiency (MMSE scores) as our patients group [13, 14]. Joussain et al. [10] obtained different results for pleasant odors, which were judged to be less pleasant by AD patients than by controls. This divergence might be explained by differences in the methodological approaches. These authors used arating technique without a hedonic visual scale to study hedonic perception and several odors for each type of odorant (six pleasant, six unpleasant, and eight neutral odors) and not a single odor per valence as in our study. Moreover, our patients had a lower mean MMSE score (19.4±3.2) than their patient group (23.9±2.3), which could also explain the difference between the results of the two studies.

The results of the hedonic responses to the three odorants showed that controls perceived these odors as pleasant, unpleasant, and neutral, as already demonstrated in the literature [22 , 31–33]. As expected, healthy subjects were able to discriminate between the three odors according to their hedonic valence using either scaling method. The AD patients discriminated between the three odors only when using the linear scale. With the five-point category scale, they perceived a significant difference between pleasant and unpleasant odors, whereas the neutral odor was simultaneously evaluated as pleasant and unpleasant. For AD patients, OCT appeared to be perceived as a perfectly neutral odor with very similar scores on both the linear and category scales. However, both BENZ and BUT were rated as more neutral on the category scale relative to the linear scale, showing that the category scale was less discriminant than the linear scale for AD patients. Nevertheless, the rank order of the mean ratings for the three odors was virtually identical between the two scales and groups. There are two possible ways to interpret the data for the hedonic scores. One is that AD patients perceive the odors with similar hedonic valence as the controls, but they use the category scale differently (but the linear scale similarly). Another way to interpret the data is that AD patients perceive the odors differently than controls for the same hedonic valence, but that only the category scale can detect the difference between the two groups of participants. The first interpretation seems to be more plausible. Indeed, several previous studies have shown that there is a general tendency among subjects to not use the extremes of the category scale (i.e., “end effects”), limiting its ability to discriminate between very well-liked or very disliked stimuli [15, 16]. Here, these “end effects”, observed only for AD patients, may be due to their lower global cognitive ability relative to controls. A previous study demonstrated the difficulty of using a category scale to evaluate the hedonic responses of young and elderly healthy individuals to apple juice in which the amount of sucrose varied. For very elderly subjects with a low education level, the authors showed that the category scale wasassociated with problems of understanding the task and expressing a subjective hedonic impression on the score-sheet [35]. Here, the education level of patients and controls did not differ, but the AD patients had a lower mean MMSE score, which could partially explain their lower discriminatory ability using category scaling than controls. Barylko-Pikilna et al. [35] also demonstrated that, for the elderly, the discriminatory power of the used scaling methods (hedonic ranking and 9-point hedonic scale) are dependent on the type of the stimuli. They showed that the ranking method was more discriminating when hedonic differences among samples were small; for products with more pronounced hedonic differences, both methods revealed equal discriminatory capacity. In another study, healthy subjects (aged between 16 and 35 years old) were not able to discriminate between different orange juice samples using hedonic ranking, whereas they were able to detect differences using the nine-point hedonic scale [17]. These observations suggest that the product set must be carefully considered to ensure that samples are not too similar or too diverse when comparing the sensitivity (discriminatory capacity) of scaling methods. Here, we only used three odorants with very contrasted hedonic valence. Future studies to confirm our results will require several less highly contrasting stimuli.

Overall, our findings on the discriminatory power show that both groups of subjects were able to correctly use both scaling methods with a slight advantage for the linear scale for AD patients. The correlation between the results obtained with the two scaling methods for each group was good: individual results with both methods and both groups were highly significantly associated. Although the correlation coefficient for the AD patients was lower than that for the controls, this observation suggests relatively well-preserved hedonic perception of AD patients. Thus, the disease had not yet affected the brain areas involved in hedonic perception, such as the amygdala, anterior cingular cortex, insula, and hippocampal area [36], of our patients with mild to moderate AD.

Repeatability is another important criterion of the sensory hedonic method. We confirmed this feature for both scaling methods and for both AD patients and controls: the two methods performed similarly (the p-values of the Student’s t-tests were above 0.05 for both). The p-values show that the two sets of data (repetition 1 and repetition 2) for each sensory scale for both groups of subjects were not significantlydifferent, but it does not prove their similarity, especially because of the low number of comparisons. The p-values for healthy controls were not very high and the test appeared to be more reliable for AD patients than controls for both scaling methods. These results need to be clarified in future studies on a larger number of subjects and the reliability of the hedonic responses further tested using more odorants, which will allow for more powerful statistical analyses.

Only the healthy controls found one method, the linear scale, to be significantly easier to use than the other. However, 61% of AD patients preferred the linear scale and only 39% preferred the category scale, although the difference was not significant. Indeed, category scales were suggested to be easier to use than linear scales in a recent review concerning the sensory and consumer methods used for older adults. Their categorical nature and limited number of choices were superior to linear scales, without structure, making them more difficult to use for some older participants [37]. Our results are not in accordance with the literature and it is possible that this is due to the very small number of subjects participating in the study. Approximately 80% of controls and 40% of patients who preferred the linear scale said that it was because it offered greater freedom to express the full range of their hedonic perception. Nevertheless, some AD patients (preferring the five-point category scale) reported that the linear scale required them to think more before reporting their hedonic assessment of each odor. All the controls and 20% of the patients preferring the five-point category scale reported that it was very simple to use, because it required that one only choose the terms corresponding to their hedonic perception of the odor. Overall, the linear scale appeared to be more appropriate for both control and patient groups in terms of ease of use.

In summary, this study has added to our knowledge concerning olfactory hedonic perception in AD. We show that patients with mild to moderate AD can use hedonic scales as well as paired healthy elderly subjects. We also show that the linear scale was slightly easier to use and provided slightly better discrimination of odorants for AD patients.

This study had some limitations. First, the AD patients participating in the present study had a slight global efficiency (MMSE score higher than 15) than the controls group. Second, only three odors of contrasting hedonic valence were used, and third, only 18 AD patients were included in the experiment. Further research with a larger number of patients is needed to confirm our findings and the study ofolfactory hedonic rating should be extended to patients with more advanced AD, using several more similar odorants, to increase the statistical power of the results. It could be informative to investigate long-term reproducibility (test-retest at several days/weeks interval) of the results, although this may be difficult with a clinical population of AD patients. It would also be useful to validate our results on individuals with other neurodegenerative or psychiatric diseases that involve cognitive deficits. Validation of the linear scale used in the present study on an elderly clinical population could significantly contribute to improving their daily life. Such a scale has possible applications, for example, in food choices offered to elderly patients in retirement homes (or other institutions) to improve their nutritional status. Moreover, significant correlations between preference and frequency of consumption have been demonstrated in the clinical population [38]. Following recommendations for changes in food preparation could help many patients enjoy their food and increase their appetite.

Footnotes

ACKNOWLEDGMENTS

The authors would like to thank G. Goutte, A. Delaune, and M. Nominé for their assistance in the sensory tests and M. Naudin for careful reading of the manuscript.

Authors’ disclosures available online ().

References

Mesholam

, Moberg

, Mahr

, Doty

(1998) Olfaction in neurodegenerative disease: A meta-analysis of olfactory functioning in Alzheimer’s and Parkinson’s diseases. Arch Neurol 55, 84–90.

Wesson

, Wilson

, Nixon

(2010) Should olfactory dysfunction be used as a biomarker of Alzheimer’s disease? Expert Rev Neurother 10, 633–635.

Naudin

, Atanasova

(2014) Olfactory markers of depression and Alzheimer’s disease. Neurosci Biobehav Rev 45, 262–270.

Godoy

MDCL

, Voegels

, Pinna

, Imamura

, Farfel

(2015) Olfaction in neurologic and neurodegenerative diseases: A literature review. Int Arch Otorhinolaryngol 2, 176–179.

Doty

, Shaman

, Dann

(1984) Development of the University of Pennsylvania Smell Identification Test: A standardized microencapsulated test of olfactory function. Physiol Behav 32, 489–502.

Hummel

, Sekinger

, Wolf

, Pauli

, Kobal

(1997) “Sniffin” sticks’: Olfactory performance assessed by the combined testing of odor identification, odor discrimination and olfactory threshold. Chem Senses 22, 39–52.

Barresi

, Ciurleo

, Giacoppo

, Foti Cuzzola

, Celi

, Bramanti

, Marino

(2012) Evaluation of olfactory dysfunction in neurodegenerative diseases. J Neurol Sci 323, 16–24.

Lombion-Pouthier

, Vandel

, Nezelof

, Haffen

, Millot

J-L

(2006) Odor perception in patients with mood disorders. J Affect Disord 90, 187–191.

Atanasova

, Graux

, El Hage

, Hommet

, Camus

, Belzung

(2008) Olfaction: A potential cognitive marker of psychiatric disorders. Neurosci Biobehav Rev 7, 1315–1325.

10.

Joussain

, Bessy

, Fournel

, Ferdenzi

, Rouby

, Delphin-Combe

, Krolak-Salmon

, Bensafi

(2016) Altered affective evaluations of smells in Alzheimer’s disease. J Alzheimers Dis 2, 433–441.

11.

Atanasova

, El-Hage

, Chabanet

, Gaillard

, Belzung

, Camus

(2010) Olfactory anhedonia and negative olfactory alliesthesia in depressed patients. Psychiatry Res 176, 190–196.

12.

Naudin

, El-Hage

, Gomes

, Gaillard

, Belzung

, Atanasova

(2012) State and trait olfactory markers of major depression. PLoS One 7, e46938.

13.

Royet

, Croisile

, Williamson-Vasta

, Hibert

, Serclerat

, Guerin

(2001) Rating of different olfactory judgements in Alzheimer’s disease. Chem Senses 26, 409–417.

14.

Naudin

, Mondon

, El-Hage

, Desmidt

, Jaafari

, Belzung

, Gaillard

, Hommet

, Atanasova

(2014) Long-term odor recognition memory in unipolar major depression and Alzheimer’s disease. Psychiatry Res 3, 861–866.

15.

Lim

(2011) Hedonic scaling: A review of methods and theory. Food Qual Prefer 22, 733–747.

16.

Pimentel

, Gomes da Cruz

, Deliza

(2016) Sensory evaluation: Sensory rating and scoring methods. In Encyclopedia of Food and Health, eds. Academic Press, Oxford, pp. 744–749 .

17.

Villanueva

NDM

, Petenate

, Da Silva

MAAP

(2005) Performance of the hybrid hedonic scale as compared to the traditional hedonic, self-adjusting and ranking scales. Food Qual Prefer 16, 691–703.

18.

Lawless

, Malone

(1986) The discriminative efficiency of common scaling methods. J Sens Stud 1, 85–98.

19.

Pause

, Miranda

, Göder

, Aldenhoff

, Ferstl

(2001) Reduced olfactory performance in patients with major depression. J Psychiatr Res 35, 271–277.

20.

Schneider

, Koch

, Reske

, Kellermann

, Seiferth

, Stöcker

, Amunts

, Shah

, Habel

(2006) Interaction of negative olfactory stimulation and working memory in schizophrenia patients: Development and evaluation of a behavioral neuroimaging task. Psychiatry Res 144, 123–130.

21.

Cumming

, Matthews

, Park

(2011) Olfactory identification and preference in bipolar disorder and schizophrenia. Eur Arch Psychiatry Clin Neurosci 261, 251–259.

22.

Hudry

, Saoud

, D’Amato

, Daléry

, Royet

J-P

(2002) Ratings of different olfactory judgements in schizophrenia. Chem Senses 27, 407–416.

23.

Guinard1

J-X

(2000) Sensory and consumer testing with children. Trends Food Sci Technol 11, 273–283.

24.

Poulton

(1979) Models for biases in judging sensory magnitude. Psychol Bull 4, 777–803.

25.

Teghtsoonian

, Teghtsoonian

(1983) Consistency of individual exponents in cross-modal matching. Percept Psychophys 33, 203–214.

26.

Stillman

(1993) Context effects in judging taste intensity: A comparison of variable line and category rating methods. Percept Psychophys 54, 477–484.

27.

Barylko-Pikielna

, Matuszewska

, Jeruszka

, Kozlowska

, Brzozowska

, Roszkowski

(2004) Discriminability and appropriateness of category scaling versus ranking methods to study sensory preferences in elderly. Food Qual Prefer 15, 167–175.

28.

Lim

, Fujimaru

(2010) Evaluation of the Labeled Hedonic Scale under different experimental conditions. Food Qual Prefer 21, 521–530.

29.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

30.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

31.

Alaoui-Ismaili

, Vernet-Maury

, Dittmar

, Delhomme

, Chanel

(1997) Odor hedonics: Connection with emotional response estimated by autonomic parameters. Chem Senses 22, 237–248.

32.

Royet

, Plailly

, Delon-Martin

, Kareken

, Segebarth

(2003) fMRI of emotional responses to odors: Influence of hedonic valence and judgment, handedness, and gender. Neuroimage 20, 713–728.

33.

Naudin

, Carl

, Surguladze

, Guillen

, Gaillard

, Belzung

, El-Hage

, Atanasova

(2014) Perceptive biases in major depressive episode. PLoS One 9, e86832.

34.

Laing

(1983) Natural sniffing gives optimum odour perception for humans. Perception 12, 99–117.

35.

Barylko-Pikielna

, Matuszewska

, Kozlowska

, Jeruszka

, Roszkowski

(2002) Perception of sweetness and sourness in apple juice varying in sucrose level and its relation to hedonic response in elderly and young adults. Pol J Food Nutr Sci 2, 65–74.

36.

Sorokowska

, Negoias

, Härtwig

, Gerber

, Iannilli

, Warr

, Hummel

(2016) Differences in the central-nervous processing of olfactory stimuli according to their hedonic and arousal characteristics. Neuroscience 324, 62–68.

37.

Methven

, Jiménez-Pranteda

, Lawlor

(2016) Sensory and consumer science methods used with older adults: A review of current methods and recommendations for the future. Food Qual Prefer 48, 333–344.

38.

Drewnowski

, Hann

, Henderson

, Gorenflo

(2000) Both food preferences and food frequency scores predict fat intakes of women with breast cancer. J Am Diet Assoc 100, 1325–1333.