An odor identification approach based on event

INTPSY-10956; No of Pages 9 International Journal of Psychophysiology xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

International Journal of Psychophysiology journal homepage: www.elsevier.com/locate/ijpsycho

An odor identification approach based on event-related pupil dilation and gaze focus Nadia Aguillon-Hernandez a,⁎, Marine Naudin b, Laëtitia Roché a, Frédérique Bonnet-Brilhault a,c, Catherine Belzung b, Joëlle Martineau a, Boriana Atanasova b a b c

Team 1 “Autism,” UMR INSERM U 930, Université François Rabelais de Tours, Tours, France Team 4 “Affective Disorders,” UMR INSERM U 930, Université François Rabelais de Tours, Tours, France CHRU de Tours, Centre Universitaire de Pédopsychiatrie, France

a r t i c l e

i n f o

Article history: Received 19 November 2014 Received in revised form 23 March 2015 Accepted 24 March 2015 Available online xxxx Keywords: Olfaction Odor identification Eye tracking Pupillary response Visual attention Gaze focus Event-related pupil dilation

a b s t r a c t Olfactory disorders constitute a potential marker of many diseases and are considered valuable clues to the diagnosis and evaluation of progression for many disorders. The most commonly used test for the evaluation of impairments of olfactory identification requires the active participation of the subject, who must select the correct name of the perceived odor from a list. An alternative method is required because speech may be impaired or not yet learned in many patients. As odor identification is known to be facilitated by searching for visual clues, we aimed to develop an objective, vision-based approach for the evaluation of odor identification. We used an eye tracking method to quantify pupillary and ocular responses during the simultaneous presentation of olfactory and visual stimuli, in 39 healthy participants aged from 19 to 77 years. Odor presentation triggered an increase in pupil dilation and gaze focus on the picture corresponding to the odor presented. These results suggest that odorant stimuli increase recruitment of the sympathetic system (as demonstrated by the reactivity of the pupil) and draw attention to the visual clue. These results validate the objectivity of this method. © 2015 Published by Elsevier B.V.

1. Introduction Alterations to olfaction have been described in various human neurodegenerative disorders (Atanasova et al., 2011), such as Alzheimer's disease (Warner et al., 1986) and Parkinson's disease (Ansari and Johnson, 1975; Doty et al., 1988b; Hawkes et al., 1997; Hawkes and Shephard, 1993; Ward et al., 1983), neurodevelopmental disorders, such as autism (Baranek et al., 2006; Kientz and Dunn, 1997; Rogers et al., 2003), and neuropsychiatric disorders, such as major depression (Pause et al., 2001; Croy et al., 2014) and schizophrenia (Moberg et al., 1997). In neurodegenerative diseases, both the peripheral and central components of the olfactory system are affected (Braak et al., 2002; Del Tredici et al., 2002; Mesholam et al., 1998; Price et al., 1991). At the very early stages, these neurodegenerative diseases can be difficult to diagnose. In particular, differential diagnosis between Alzheimer disease (AD) and depression in elderly cannot be easy to establish (McLean, 1987) and may impede early therapeutic care. However, these two categories of patients differ in the nature of the olfactory disorder, studies comparing odor identification performance between patients AD and patients with depression have reported most important ⁎ Corresponding author at: Inserm U 930, Bâtiment B1A, CHRU Bretonneau, 37044 Tours Cedex 9, France. Tel.: +33 2 47 47 97 47; fax: +33 2 47 47 67 47. E-mail address: [email protected] (N. Aguillon-Hernandez).

alterations in AD compared to patients with depression (McCaffrey et al., 2000; Pentzek et al., 2007; Solomon et al., 1998). Recent studies have shown the potential interest of olfactory tests to aid in early diagnosis of these diseases (Naudin et al., 2014). Concerning, autistic spectrum disorder, subjects have also been shown to be significantly less accurate than controls in olfactory identification tests (Bennetto et al., 2007). These findings may have implications for our understanding of these disorders, with the inclusion of sensorial disorders in the clinical profiles of patients, potentially facilitating diagnosis and improving patient care. The tests most widely used to investigate olfactory identification disorders in clinical research are the University of Pennsylvania Smell Identification Test (UPSIT) (Doty et al., 1984) and the “Sniffin' Sticks” test (Hummel et al., 1997). In these tests, the participants are asked to identify the perceived odor, in a four-alternative forced-choice procedure. These tests are currently the best tools available for the diagnosis of olfactory disorders. However, they are subject to certain limitations because they require the active participation of the subject and good language acquisition. Speech may be impaired or not yet learned in some patients, so an alternative approach is required. We propose here an alternative method for evaluating odor identification that does not involve speech or active cognitive participation and is therefore suitable for use with a large number of patients. This alternative objective method is based on the principle of the multimodal

http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009 0167-8760/© 2015 Published by Elsevier B.V.

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

2

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx

convergence of primary sensory inputs. Studies in mammals (rat and monkey) suggest that primary sensory information converges on common structures, such as the hippocampus (Deadwyler et al., 1987) and orbitofrontal cortex (Carmichael and Price, 1995). The orbitofrontal cortex has been shown to receive afferent input from both the primary olfactory (piriform) cortex and visual association areas (Carmichael and Price, 1995), and as has been shown in humans by an fMRI study, it participates in the orientation of visual attention in the presence of a salient visual stimulus (Armony and Dolan, 2002). In our study, we call this orientation of the eyes guided by exogenous events: the eventrelated gaze focus (ERGF). Multimodal integration of this type has been described in studies combining auditory and visual stimuli. For example, subjects hearing a word spontaneously oriented their gaze to the picture corresponding to this word in an eye tracking study (Huettig and Altmann, 2005). A similar phenomenon has been demonstrated in studies of the multimodal integration of visual and olfactory stimuli, in which odors were found to draw the attention of test subjects to visual objects corresponding to the odors concerned (Chen et al., 2013; Seigneuric et al., 2010; Seo et al., 2010). The orbitofrontal cortex is involved in the orientation of visual attention in the presence of a stimulus, and odor identification is facilitated by searches for visual clues (Chen et al., 2013; Gottfried and Dolan, 2003; Seigneuric et al., 2010; Seo et al., 2010). We therefore used this principle to develop a method based on perceptual processes without the need for active cognitive processes, for tests of odor identification in patients unable to participate in conventional tests. We also assessed subjective perceptions of the intensity and hedonism of the odor, as such factors have been shown to influence odor identification (Kare, 2012). In addition to the ERGF, we also considered a physiological response to olfactory stimulation—pupil dilation (Winneke, 1992)— which reflects the physiological mobilization of the body. Various autonomic parameters, including heart rate (Alaoui-Ismaïli et al., 1997; Bensafi et al., 2002a; Brauchli et al., 1995), electrodermal activity (Bensafi et al., 2002b; Borsanyi and Blanchard, 1962; Møller and Dijksterhuis, 2003; Shock and Coombs, 1973), respiratory frequency (Doty et al., 1988a; Eccles et al., 1989; Laing, 1983) and pupil size (Schneider et al., 2009; Winneke, 1992), react to olfactory stimuli. Pupillary response is known to be modulated by two major reflexes: the photomotor reflex (inducing constriction of the pupil in response to light stimulation under the influence of the parasympathetic nervous system) (Beatty and Lucero-Wagoner, 2000) and the psychosensory reflex (inducing dilation of the pupil in response to a cognitive load, under the influence of the sympathetic nervous system) (Andreassi, 2000; Granholm et al., 1996). Pupillary dilation is a useful parameter for two key reasons: (1) changes in pupillary dilation are more sensitive than many other physiological parameters (Bradley et al., 2008; Kahneman and Peavler, 1969), and (2) this parameter can be recorded simultaneously with gaze position with a non-contact method (Martineau et al., 2011). In our study, we call this pupil dilation in response to stimulation: the event-related pupil dilation (ERPD). The aim of our study was to validate, in an adult population, the association between odor recognition and the direction of attention to a picture corresponding to the odor and pupillary dilation. Using an eye tracking system to measure the pupil (ERPD) and visual exploratory responses (ERGF) associated with olfactory stimulation, we tested the hypothesis that olfactory stimulation leads to pupil dilation and an increase in exploration of the picture corresponding to the odor. 2. Experimental procedures

(urban or countryside). They were provided with full details about the experimental protocol before the tests began. All participants were informed of their rights before taking part in procedures approved by the local ethics committee, and all gave written informed consent in accordance with the Helsinki Declaration (World, 2004). The inclusion criteria included normal or corrected vision, no history of an eye disorder, and no self-reported problems relating to sense of smell. The exclusion criteria for all participants included possible brain damage, major medical problems, intake of pharmacological molecules targeting the vegetative system, current substance abuse, allergy, a current cold, or a transient problem with sense of smell. All subjects were selected on the basis of an absence of anosmia to the odorants used. 2.2. Stimuli 2.2.1. Olfactory stimuli We presented 10 familiar odors, corresponding to plants (mint), foods (orange, mushroom, bread, vanilla, coconut, strawberry, raw potato), industrial materials (oil paint), and agricultural products (wet earth). The olfactory stimuli used, developed by Sentosphere®, were non-toxic and were presented in a solid form, in a cup at suprathreshold concentrations. They were obtained by adding between 10% and 30% (according to the odor) of fragrance ingredient on the “seeds”, which allow fixing the odors on a solid support. All olfactory stimuli were presented during 10 s. 2.2.2. Visual stimuli Olfactory stimuli were presented simultaneously with a visual stimulus consisting of a board bearing four color pictures. The creation of the database of visual stimuli involved three stages of validation. 1) All the pictures on a given board belonged to the same semantic field (e.g., forest smells). Picture fidelity was validated with a questionnaire completed by 150 people aged from 2 to 80 years (mean age = 37.6 years ± 15.1), and only pictures accepted by over 80% of respondents were used to form the database. (2) On each board, the surface brightness and visual characteristics of each picture were homogenized and a pilot study was carried out with 15 people (mean age = 28.8 ± 12.6 years) to check that each image had a similar influence on visual exploration behavior in the absence of olfactory stimulation. (3) The four pictures on each board included one corresponding to the odor (the target), a picture of an item with a smell very similar to that presented (the competitor. For example, for the smell of orange, the competitor image was a picture of a lemon), a picture of an item with a smell different from that presented (distractor 1; for example, for the smell of orange, the distractor 1 image was that of a kiwi) and a picture of an item with a smell very different from that presented (distractor 2; for example, for the smell of orange, the distractor 2 image was that of a pear). This gradient of correspondence between the items and the odor was established in a pilot study including 10 control participants (mean age = 31.1 years ± 4.9). Between boards, the participants were presented with an image of a central cross, to refocus the gaze. The pupillary reflex in response to light (photomotor reflex) was prevented by ensuring that the brightness of the cross image was identical to that of the boards. The stimuli were presented on a computer screen (17″), at a distance of about 90 cm from the subjects' eyes. Each board was 47.5 cm × 27 cm in size, with a resolution of 1920 × 1080 px (visual angle = 30°). All visual stimuli were homogenized (brightness (20 lx), color (histogram values for the red, green, and blue components of each board: R = 242, G = 239, B = 234), and size (1920 × 1080 px)).

2.1. Participants 2.3. Materials We recruited 39 non-smokers (23 women and 16 men), aged 19 to 77 years (mean ± SD age = 37.6 ± 15.1 years). At recruitment, the subjects were informed that they would have to smell different compounds that can be found in food products and in the environment

2.3.1. Eye tracking Visual stimuli were delivered by the head-free mounted FaceLab® eye tracking system, which consists of a computer equipped with two

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx

digital infrared light cameras and an infrared light source (wavelength 875 nm) derived from the international exposure standards set by the International Electrotechnical Commission (IEC., 2001). The subject was not directly fitted with any equipment, and the corneal reflection of infrared light was measured to monitor ocular behavior and pupil size variation. 2.3.2. Delivery of olfactory stimuli The experiment was conducted in a blind room with an automatic permanent ventilation (Fig. 1a).The experimenter positioned the odor during approximately 10 s just below the lower lip of the subject, to ensure that eye tracking was not disturbed. All participants breathed in spontaneously from the start of olfactory stimulation. The experimenter was sat next to the participant, slightly back so as not to invade his visual field (Fig. 1a). The experimenter had a timed gauge on the control station allowing him to prepare for the smell position near the face of the subject. He had the olfactory cup on a small rigid device Plexiglas (length: 40 cm, width: 3 cm, thickness: 5 mm) (Fig. 1b). The experimenter controlled his gesture from his monitor so as not to invade the visual field of the participant and not to interfere with the tracking. 2.4. Experimental design Recordings were made in three stages: a stage without olfactory stimulation (Non-olfactory run) and two stages with olfactory stimulation (Olfactory objective run and Olfactory subjective run). The “Objective olfactory run” always precedes the “Subjective olfactory run” to avoid the effect of the instruction given during the “Subjective olfactory run” (i.e., “verbally identify odors”). During the “Objective olfactory run”, the goal is to measure a spontaneous response to the olfactory stimulation whereas in the “Subjective olfactory run”, the objective is to measure only the identification of odors. 2.4.1. Non-olfactory run During the stage without olfactory stimulation, only visual stimuli were presented, for 6 s each, in a random order. The visual stimuli were preceded by the image of a central cross, which was presented for 4 s. During this run, the only instruction issued to the subjects was to pay attention to the board. During this run, pupil diameter was measured during presentation of the cross, and exploratory parameters were measured during presentation of the board (see below for more details concerning these measurements) (Fig. 2a). This “Non-olfactory

3

run” was preceded by the “Objective olfactory run” for half the subjects, and followed by this run for the other half.

2.4.2. The objective olfactory run The paradigm for the “Objective olfactory run” was similar to that described above except that an olfactory stimulus was presented from the beginning of presentation of the cross until the end of the presentation of the board. Olfactory stimulation lasted 10 s and was followed by evaluations on two categorical scales, to evaluate the odor's hedonic aspect (the participant was asked to complete the statement “I like that smell…” with one of the following responses: “not at all”, “not much”, “a little”, “moderately”, or “a lot”) and intensity (the participant was asked to complete the statement “the smell is…” with one of the following responses: “very weak”, “weak”, “moderately strong”, “strong”, or “very strong”), with 30 s for recovery time left between olfactory stimuli. During this run, the only instruction given to the subjects was to pay attention to the odor and the board for the assignment of hedonic and intensity values. Pupil diameter and ocular parameters were also measured during this run (Fig. 2b).

2.4.3. The subjective olfactory run The “Subjective olfactory run” was carried out last in the protocol. It closely resembled the “Objective olfactory run” except for the order of presentation of odors. During this run, participants were asked to identify the presented odors verbally, possibly with the assistance of the visual stimuli. In order to avoid the participation of the subjects with anosmia, all subjects that could not perceive one or more odors were eliminated. Moreover, during the measurement sessions (subjective run), all participants were able to perceive 100% of odors presented. We checked that there was no effect of the repetition of odor presentation (once during the “Objective olfactory run” and then again during the “Subjective olfactory run”) in a pilot study on 20 control participants aged 20 to 60 years (mean age = 43.2 ± 12.1 years). In this pilot study, 10 participants underwent objective olfactory testing after a control run (visual stimuli only) and subjective olfactory testing (conditions identical to those used for the experiment, with participants being subjected twice to each odor). The other 10 participants underwent objective olfactory testing after the control run only (i.e., no subjective olfactory testing; participants subjected only once to each odor). We found that odor repetition had no effect on the subjective identification of odors (Mann–Whitney analysis: U = 48.5; p = 0.93) (Fig. 2c).

Fig. 1. Presentation of olfactory stimulation. The experimenter was sat next to the participant and had a timed gauge on the control station allowing him to prepare for the smell position near the face of the subject and to control duration of the olfactory stimulation (approximately10 s). The experimenter had to set the olfactory cup on a small rigid device Plexiglas (length: 40 cm, width: 3 cm, thickness: 5 mm). The experimenter controlled his gesture from his monitor so as not to invade the visual field of the participant and not to interfere with the tracking. The removable separation prevents the participant to see the control station.

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

4

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx

2.5. Measurements and statistical analysis 2.5.1. Objective data We checked that all exploratory parameters and pupil data were normally distributed by carrying out Kolmogorov–Smirnov analysis. The homogeneity of variances was checked by Levene analysis. 2.5.1.1. Event-related pupil dilation (ERPD). Measurements. Pupillary diameter was measured every 100 ms during presentation of the cross image for the “Non-olfactory run” and the “Objective olfactory run”. Pupil size variation was assessed with respect to baseline pupil diameter (measured 500 ms before the start of stimulation). Pupil diameter variation was calculated by taking the difference between the baseline and the mean change in pupil size between 1 s and 4 s (during the presentation of the cross). Pupillary reactivity has a latency of 1 s (peak pupillary reactivity (Beatty, 1982; Beatty and Lucero-Wagoner, 2000)); we have excluded the values of the pupil recorded during the first second. Statistical analysis. As in all studies analyzing the parameters related to the event (for example, as evoked potentials; Gonsalvez and Polich, 2002), only the responses corresponding to the event are studied. Thus pupil size variation was studied by measuring the response to olfactory stimulation for each subject only if the odor was identified correctly during the “Subjective olfactory run” (i.e., if a participant correctly identified 8 of the 10 odors presented during the “Subjective olfactory run”, only the responses to these eight odors were taken into account in the analysis). The effect of the test conditions (olfactory conditions during the “Subjective olfactory run” vs. nonolfactory conditions during the “Non-olfactory run”) on pupil variation was assessed in a GLM analysis, including sex (male vs. female) as a categorical predictor and age as a continuous predictor. We then carried out Dunn–Bonferroni post hoc analysis to compare paired means. 2.5.1.2. Event-related gaze focus (ERGF). measurements. Ocular strategy was quantified by measuring the position of the eye with respect to various parameters of the board carrying four images. Regions of interest (ROI) (of identical size except for the center) were established on the target, the competitor, and the two distractors and the center of the board (Fig. 3). The time spent exploring and fixating on each region of interest was recorded during the “Non-olfactory run” and the “Objective olfactory run”. Gaze fixation time (correspond to the fixing of the gaze) was detected with a dispersion based-algorithm, according to which, points of focus tend to cluster closely together because of the low velocity of eye movement when the gaze is fixed (Salvucci and Goldberg, 2000). Fixation on a particular area was identified by considering an unchanged position of the eye (3 gaze points less than 40 pixels apart, corresponding to a visual angle of 1°, giving a foveal angle of 2°) for at least 50 ms to correspond to a point on which the gaze was fixed (as in the procedure defined by Dalton et al., 2005. The number of gaze fixations (number of times the gaze focused on an ROI) and the latency of the first gaze fixation (the time to first entry of the gaze into each ROI) were also calculated during the “Nonolfactory run” and the “Objective olfactory run”.

5

exploratory parameters was assessed for each ROI by GLM analysis, including conditions and ROI as nested effects (level 1: [“Non-olfactory run” vs. “Objective olfactory run”]; level 2: [target, competitor, distractor 1, and distractor 2]), sex as a categorical predictor (male vs. female) and age as a continuous predictor. The analysis was corrected with the Greenhouse–Geiser test and the Dunn–Bonferroni post hoc analysis for the comparison of paired means. 2.5.2. Subjective data 2.5.2.1. Odor identification: measurements. During the “Subjective olfactory run”, participants were asked to identify the odor presented with the assistance of the images presented. The participants' responses were used to calculate an identification score for each odor. A binary score was used for the evaluation of odor identification. A score of 2 was attributed if the participant correctly identified the odor. A score of 1 was assigned if the odor was not correctly identified. A mean score was then calculated for the identification of each of the 10 odors. 2.5.2.2. Hedonic evaluation: measurements. During the “Objective olfactory run”, a subjective assessment of the hedonic perception of the odor was obtained through the attribution of a score on a hedonism scale. For future application of this method in the clinical population (sometimes having cognitive problems), we have chosen a rating scale of 5 points. We assigned scores as follows: 1 for “not at all”, 2 for “not much”, 3 to “moderately”, 4 to “a little”, and 5 to “a lot”. A hedonic score was thus calculated for each odor from the responses given by the participants. Odors were considered pleasant if they scored more than 3 on this scale, and unpleasant if they scored less than 3. No odor was considered neutral (score strictly equal to 3). 2.5.2.3. Intensity evaluation: measurements. During the “Objective olfactory run”, the intensity of the odor was assessed in a subjective manner, by the assignment of a score on an intensity scale. Scores were assigned as follows: 1 for “very weak”, 2 for “weak”, 3 for “moderately strong”, 4 for “strong”, and 5 for “very strong. An intensity score was thus calculated for each odor, on the basis of the responses given by the participants. Odors receiving scores of 4 or more were considered to be of high intensity, those with a score of 3 were considered to be of medium intensity and those with scores of 2 or below were considered to be of low intensity. 2.5.3 . Correlation between objective and subjective data Correlations between objective data (exploratory parameters and pupil variation) and subjective scores (identification score, hedonic score, and intensity score) were assessed in Spearman's rank correlation test. Concerning the objective data, all responses were included (even for unsuccessful identifications of the odor during the “Subjective olfactory run”) because the identification scores during subjective run considers both correct and incorrect responses 3. Results

2.5.1.3. Statistical analysis. ERGF parameters were analyzed by measuring the response to olfactory stimuli for each subject only for odors identified correctly in the “Subjective olfactory run”. The effect of the test conditions (olfactory conditions during the “Objective olfactory run” vs. non-olfactory conditions during the “Non-olfactory run”) on

3.1. Objective data No effect of sex or age was identified for any of the parameters considered (p N 0.05).

Fig. 2. Experimental paradigm. The “Non-olfactory run” and The “Objective olfactory run” were carried out in random order, with the “Subjective olfactory run” always carried out last. Each run began with a white board (2 s) followed by black board (2 s) (to ensure an intact pupillary response to light) and each run consisted of 10 sequences (one sequence per odor). Each sequence began with a board with a cross in the center (4 s) followed by a board with 4 pictures. Then in the “Objective olfactory run” (b) and in the “Subjective olfactory run” (c) a board with a hedonism scale (15 s) and a board with an intensity scale (15 s) were presented (but these 2 boards were only exploited during the “Objective olfactory run”). Pupil dilation was recorded in response to presentation of the cross board without olfactory stimulation or with olfactory stimulation. Visual exploration parameters were recorded in response to the picture-board presentation without (a) or with olfactory stimulation (b). Subjective evaluation of hedonism (hedonic score) and odor intensity (intensity score) were recorded after the picture-board presentation with olfactory stimulation (b). Objective odor identification was recorded in response to the picture-board presentation with olfactory stimulation (c).

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

6

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx

Fig. 3. Region of interest (ROI): 1 central ROI, 1 target ROI, 1 competitor ROI, and 2 distractor ROIs.

3.1.1. ERPD Mean pupil size at the start of the experiment was 4.01 mm. Olfactory stimulation during the “Objective olfactory run” was accompanied by a 0.29 mm increase in pupil size (a 7.25% increase; F1,36 = 18.42; p b 0.001). The absence of olfactory stimulation during the “Non-olfactory run” was accompanied by a decrease in pupil size to below the baseline value (-0.43 mm, corresponding to a 10.75% decrease; Fig. 4).

3.1.4. Number of gaze fixation on ROIs During the “Objective olfactory run”, olfactory stimulation affected the number of times that participants focused on the different ROIs (F3,108 = 5.35; p = 0.004): participants looked at the target a significantly larger number of times than they looked at the competitor (p b 0.0001), and the two distractors (p b 0.0001; p b 0.0001) during the “Objective olfactory run”. They looked at the target a significantly larger number of times during the objective run than during the control run (p b 0.0001) (Fig. 5c).

3.1.2. ERGF During the “Objective olfactory run”, olfactory stimulation affected the exploration of the different ROIs (F3,108 = 10.96; p = 0.001): participants spent significantly longer exploring the target than the competitor (p b 0.0001) or the two distractors (p b 0.0001 and p b 0.0001, respectively) during the “Objective olfactory run”. Participants spent longer exploring the target during the objective olfactory run than during the control run (p b 0.0001) (Fig. 5a).

3.1.5. Latency of entry into an ROI During the “Objective olfactory run”, olfactory stimulation affected the time to entry of the gaze into an ROI (F1,108 = 8.58; p b 0.0001): participants directed their gaze to the target ROI significantly more rapidly than to the competitor ROI (p = 0.01) or distractor ROIs (p b 0.001 for distractor 1; p b 0.0001 for distractor 2); they also directed their gaze to the target ROI significantly more rapidly during the objective olfactory run than during the control run (p = 0.002) (Fig. 5d).

3.1.3. Gaze fixation time During the “Objective olfactory run”, olfactory stimulation affected the time spent focusing on the different ROIs (F3,108 = 11.8; p b 0.0001): significantly longer was spent looking at the target than at the competitor (p b 0.0001) and the two distractors (p b 0.0001; p b 0.0001) during the “Objective olfactory run”. Participants spent longer looking at the target during the “Objective olfactory run” than during the “Non-olfactory run” (p b 0.0001) (Fig. 5b).

3.1.6. Correlation with subjective data Identification score was positively correlated with time spent exploring the target during “Objective olfactory Run” (r = 0.82; p = 0.004): the time spent exploring the target increased with identification score (Fig. 6d). Pupil dilation was found to be positively correlated with intensity score (r = 0.73; p = 0.02): higher intensity scores were associated with greater dilation of the pupil (Fig. 6a). Identification score was positively correlated with intensity score (r = 0.63; p = 0.05; Fig. 5b), and hedonism score was positively correlated with intensity score (r = 0.64; p = 0.04; Fig. 6c). 4. Discussion

Fig. 4. Pupil response. Mean pupil diameter variation during presentation of the cross board in the control and objective olfactory runs. Significantly greater variation in response to olfactory stimulation was observed in the objective run than in the control run. *p b 0.05, **p b 0.01, ***p b 0.001.

The aim of this work was to determine whether ocular responses could be used to assess odor identification in an objective manner. We found that olfactory stimulation induced dilation of the pupil and a rapid focus of the gaze toward the visual clue corresponding to the odor. Our study suggests that ERGF and ERPD could be used as an objective indicator of odor perception and identification. We first showed that olfactory stimulation induced pupillary dilation, confirming the results of a previous study (Schneider et al., 2009). Pupillary activity is classically associated with mental effort (Hess and Polt, 1960; Hyönä et al., 1995; Kahneman and Peavler, 1969). ERPD seems to reflect the mobilization of attentional resources, as demonstrated by the strong link between the activity of the reticular system (especially the locus caeruleus, which modulates arousal and attention (Aston-Jones et al., 1999)) and pupil activity (Murphy et al., 2014). The ERPD is thus a very interesting parameter, providing information about the activity of the autonomic nervous system (Steinhauer et al., 2004). Olfactory stimulation induces an increase in autonomic activity via reticular formation recruitment (Motokizawa, 1974), thereby modulating sympathetic activity (Hayes and Weaver, 1992). Olfactory stimulation may therefore increase

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx

7

Fig. 5. Exploratory parameters. Exploratory parameters recorded during the “Objective olfactory run” (black) and the “Non-olfactory run” (gray). (a) Time (s) spent exploring ROIs: olfactory stimulation resulted in a larger amount of time being spent exploring the target (p b 0.05, **p b 0.01, ***p b 0.001). (b) Gaze fixation time (s) on ROI: olfactory stimulation resulted in an increased gaze fixation time on the target (p b 0.05, **p b 0.01, ***p b 0.001). (c) Number of gaze fixation on an ROI: olfactory stimulation resulted in an increased of number of gaze fixation on the target (p b 0.05, **p b 0.01, ***p b 0.001). (d) Latency of entry (s) into ROI: olfactory stimulation decreased the time to entry of the gaze into the target ROI (p b 0.05, **p b 0.01, ***p b 0.001).

recruitment of the sympathetic system, resulting in pupillary dilation, whereas an absence of olfactory stimulation was found to decrease pupil dilation to below baseline levels. This may result from an attentional effect: the absence of a stimulating event during presentation of the cross board during the “Non-olfactory run” may have resulted in a decrease in attention, leading to a decrease in pupil size. Our data also highlighted the existence of a relationship between pupil dilation and estimates of stimulus intensity. Indeed, more intense stimuli were associated with greater pupil dilation. Presumably, olfactory stimulation induces changes in a physiological parameter under autonomic control,

making it possible to determine whether the stimulation is felt strongly or not. This aspect is potentially useful for objective studies of olfactory detection thresholds or irritation thresholds corresponding to excessively high intensities of stimulation. These findings are similar to those reported by Schneider et al. (2009), who established a link between pupil dilation and odor intensity with stimuli at different concentrations. We demonstrated a link between subjective assessments of the intensity of stimulation and pupil reactivity. Presumably, assessments of the intensity of odors may be influenced by the subjective experience of participants. The relationship between the hedonistic

Fig. 6. Correlations. (a) Correlation between pupil dilation (mm) and intensity score (1 corresponds to a very low intensity and 5 to a very strong intensity). (b) Correlation between identification score (1 corresponds to an incorrect identification and 2 to a correct identification) and intensity score (1 corresponds to a very low intensity and 5 to a very strong intensity). (c) Correlation between hedonism score (1 corresponds to a very unpleasant odor and 5 to a very pleasant odor) and intensity score (1 corresponds to a very low intensity and 5 to a very strong intensity).

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

8

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx

score and intensity estimates for odors observed in this study is consistent with this argument. This result is consistent with previous studies demonstrating an association (linear or nonlinear) between the intensity of stimulation and hedonistic value (Kare, 2012), although a dissociation has been reported for neural representations of intensity and valence (Anderson et al., 2003). Our results indicate that variations in pupil size can be used as an objective index of the subjective perception of olfactory stimulation and that this reactivity could be a useful tool for the objective assessment of olfactory perception, without the need for active participation by the patient, in many diseases. We also found a link between intensity and identification score, suggesting that odors were better identified when they were perceived more strongly. Our study shows that the amplitude of the ERPD is correlated with the intensity and hedonic value of the stimulus. Our findings also indicate that visual exploratory behavior may be influenced by odor identification. All visual scanning parameters were influenced by olfactory stimulation, with convergence on the image corresponding to the odor (when participants correctly identified the odor). Odor identification was associated with a longer time spent exploring and looking at the target, and a shorter time to the initiation of target exploration. These data are consistent with those of Seo et al. (2010), which provided the first demonstration of an olfactory priming effect on visual selective attention. We also provide additional data concerning time to exploration of the image corresponding to the odor presented (Seo et al., 2010). The greater time spent exploring and looking at the target and the shorter time to exploration of the target image indicates a greater focusing of attention on the target. The positive correlation between the time spent exploring the target and identification score also suggests that attention is directed to the image when the odor is correctly identified. The shorter time to beginning exploration of the target again suggests that the target captures the attention. These data confirm previous studies recording eye movements associated with olfactory stimulation during the exploration of a complex scene (including a picture of an odorant object) (Seigneuric et al., 2010). The short time taken for participants to direct their gaze at the target ROI may reflect the early capture of visual attention, mediated by automatic processes. Chen et al. (2013) showed that the presentation of visual objects with olfactory stimulation increased perceptual saliency. We demonstrate here that olfactory stimulation induces gaze focus response and physiological responses. These responses are potentially interesting tools for the investigation of odor identification in populations unable to participate in conventional olfactory tests. Furthermore, of the use of this approach to study diseases, including disorders of olfaction, might lead to the detection of potential biomarkers. Gaze and pupil responses can be a complementary exploration of olfactory disorders, especially in difficult clinical populations with cognitive disorders (or very young patients with a low level of development, for example). Eye tracking solution is interesting and easily applicable because recent system are more compact and mobile (e.g., Tobii X230, X2-60 Tobii, Tobii Glasses 2, SMI-RED250mobile, SMI-REDN, SMI Eye Tracking Glasses 2 Wireless), and some are very low cost but needs to develop oneself analysis software (e.g., Eye Tribe for $99). The new eye tracking systems are becoming easier to use; however, it is necessary to remain vigilant and to make it a good practice. In particular, the quality of visual stimuli must be rigorous and light conditions must be controlled. In conclusion, this study conclusively demonstrates the validity of our objective and implicit approach to the evaluation of odor identification. Pupil dilation appears to be an objective indicator of physiological response to olfactory stimulation. The longer periods of time spent looking at the target suggest that odor stimulation directs visual attention towards the target. This validates our assumptions about the objectivity of this method. The next stage of this work will be to conduct a cross-sectional study in young participants and patients.

References Alaoui-Ismaïli, O., Robin, O., Rada, H., Dittmar, A., Vernet-Maury, E., 1997. Basic emotions evoked by odorants: comparison between autonomic responses and self-evaluation. Physiol. Behav. 62, 713–720. Anderson, A.K., Christoff, K., Stappen, I., Panitz, D., Ghahremani, D.G., Glover, G., Gabrieli, J.D.E., Sobel, N., 2003. Dissociated neural representations of intensity and valence in human olfaction. Nat. Neurosci. 6, 196–202. Andreassi, J.L., 2000. Psychophysiology: human behavior and physiological response. Lawrence Erlbaum, Mahwah, N.J., p. 550. Ansari, K.A., Johnson, A., 1975. Olfactory function in patients with Parkinson's disease. J. Chronic Dis. 28, 493–497. Armony, J.L., Dolan, R.J., 2002. Modulation of spatial attention by fear-conditioned stimuli: an event-related fMRI study. Neuropsychologia 40, 817–826. Aston-Jones, G., Rajkowski, J., Cohen, J., 1999. Role of locus coeruleus in attention and behavioral flexibility. Biol. Psychiatry 46, 1309–1320. Atanasova, B., Hernandez, N., Nieuwenhuijzen, P., Mondon, K., Belzung, C., 2011. Psychophysical, neurophysiological and neurobiological investigation of olfactory process in humans. In: Weiss, L.E., Atwood, J.M. (Eds.), The Biology of Odors: Sources, Olfaction and Response, pp. 1–68. Baranek, G.T., David, F.J., Poe, M.D., Stone, W.L., Watson, L.R., 2006. Sensory Experiences Questionnaire: discriminating sensory features in young children with autism, developmental delays, and typical development. J. Child Psychol. Psychiatry 47, 591–601. Beatty, J., 1982. Task-evoked pupillary responses, processing load, and the structure of processing resources. Psychol. Bull. 91, 276. Beatty, J., Lucero-Wagoner, B., 2000. The pupillary system. Handb. Psychophysiol. 2, 142–162. Bennetto, L., Kuschner, E.S., Hyman, S.L., 2007. Olfaction and taste processing in autism. Biol. Psychiatry 62, 1015–1021. Bensafi, M., Rouby, C., Farget, V., Bertrand, B., Vigouroux, M., Holley, A., 2002a. Influence of affective and cognitive judgments on autonomic parameters during inhalation of pleasant and unpleasant odors in humans. Neurosci. Lett. 319, 162–166. Bensafi, M., Rouby, C., Farget, V., Bertrand, B., Vigouroux, M., Holley, A., 2002b. Autonomic nervous system responses to odours: the role of pleasantness and arousal. Chem. Senses 27, 703–709. Borsanyi, S.J., Blanchard, C.L., 1962. Psychogalvanic skin response olfactometry. Ann. Otol. Rhinol. Laryngol. 71, 213. Braak, H., Del Tredici, K., Bratzke, H., Hamm-Clement, J., Sandmann-Keil, D., R\üb, U., 2002. Staging of the intracerebral inclusion body pathology associated with idiopathic Parkinson's disease (preclinical and clinical stages). J. Neurol. 249, III-1. Bradley, M.M., Miccoli, L., Escrig, M.A., Lang, P.J., 2008. The pupil as a measure of emotional arousal and autonomic activation. Psychophysiology 45, 602–607. Brauchli, P., Rüegg, P.B., Etzweiler, F., Zeier, H., 1995. Electrocortical and autonomic alteration by administration of a pleasant and an unpleasant odor. Chem. Senses 20, 505–515. Carmichael, S.T., Price, J.L., 1995. Sensory and premotor connections of the orbital and medial prefrontal cortex of macaque monkeys. J. Comp. Neurol. 363, 642–664. Chen, K., Zhou, B., Chen, S., He, S., Zhou, W., 2013. Olfaction spontaneously highlights visual saliency map. Proc. R. Soc. B Biol. Sci. 280, 20131729. Croy, I., Symmank, A., Schellong, J., Hummel, C., Gerber, J., Joraschky, P., Hummel, T., 2014. Olfaction as a marker for depression in humans. J. Affect. Disord. 160, 80–86. Dalton, K.M., Nacewicz, B.M., Johnstone, T., Schaefer, H.S., Gernsbacher, M.A., Goldsmith, H.H., Alexander, A.L., Davidson, R.J., 2005. Gaze fixation and the neural circuitry of face processing in autism. Nat. Neurosci. 8, 519–526. Deadwyler, S.A., Foster, T.C., Hampson, R.E., 1987. Processing of sensory information in the hippocampus. CRC Crit. Rev. Clin. Neurobiol. 2, 335–355. Del Tredici, K., Rüb, U., De Vos, R.A., BOHL, J.R., BRAAK, H., 2002. Where does Parkinson disease pathology begin in the brain? J. Neuropathol. Exp. Neurol. 61, 413. Doty, R.L., Shaman, P., Kimmelman, C.P., Dann, M.S., 1984. University of Pennsylvania Smell Identification Test: a rapid quantitative olfactory function test for the clinic. Laryngoscope 94, 176–178. Doty, R.L., Deems, D.A., Frye, R.E., Pelberg, R., Shapiro, A., 1988a. Olfactory sensitivity, nasal resistance, and autonomic function in patients with multiple chemical sensitivities. Arch. Otolaryngol. Head Neck Surg. 114, 1422–1427. Doty, R.L., Deems, D.A., Stellar, S., 1988b. Olfactory dysfunction in parkinsonism: general deficit unrelated to neurologic signs, disease stage, or disease duration. Neurology 38, 1237–1244. Eccles, R., Jawad, M.S., Morris, S., 1989. Olfactory and trigeminal thresholds and nasal resistance to airflow. Acta Otolaryngol. Stockh. 108, 268–273. Gonsalvez, C.L., Polich, J., 2002. P300 amplitude is determined by target-to-target interval. Psychophysiology 39, 388–396. Gottfried, J.A., Dolan, R.J., 2003. The nose smells what the eye sees: crossmodal visual facilitation of human olfactory perception. Neuron 39, 375–386. Granholm, E., Asarnow, R.F., Sarkin, A.J., Dykes, K.L., 1996. Pupillary responses index cognitive resource limitations. Psychophysiology 33, 457–461. Hawkes, C.H., Shephard, B.C., 1993. Selective anosmia in Parkinson's disease? Lancet 341, 435–436. Hawkes, C.H., Shephard, B.C., Daniel, S.E., 1997. Olfactory dysfunction in Parkinson's disease. J. Neurol. Neurosurg. Psychiatry 62, 436–446. Hayes, K., Weaver, L.C., 1992. Tonic sympathetic excitation and vasomotor control from pontine reticular neurons. Am. J. Physiol. Heart Circ. Physiol. 263, H1567–H1575. Hess, E.H., Polt, J.M., 1960. Pupil size as related to interest value of visual stimuli. Science 132 (3423), 349–350. Huettig, F., Altmann, G.T.M., 2005. Word meaning and the control of eye fixation: semantic competitor effects and the visual world paradigm. Cognition 96, B23–B32.

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009

N. Aguillon-Hernandez et al. / International Journal of Psychophysiology xxx (2015) xxx–xxx Hummel, T., Sekinger, B., Wolf, S., Pauli, E., Kobal, G., 1997. “Sniffin' Sticks”: olfactory performance assessed by the combined testing of odor identification, odor discrimination and olfactory threshold. Chem. Senses 22, 39. Hyönä, J., Tommola, J., Alaja, A.-M., 1995. Pupil dilation as a measure of processing load in simultaneous interpretation and other language tasks. Q. J. Exp. Psychol. Sect. A 48, 598–612. IEC., 2001. Safety of laser products—Part 1: equipment classification, requirements and user's guide (IEC-60825-1 1993A1 1997A2 2001 Available from wwwiecch). Kahneman, D., Peavler, W.S., 1969. Incentive effects and pupillary changes in association learning. J. Exp. Psychol. 79, 312–318. Kare, M.R., 2012. The Chemical Senses and Nutrition. Elsevier, p. 511. Kientz, M.A., Dunn, W., 1997. A comparison of the performance of children with and without autism on the Sensory Profile. Am. J. Occup. Ther. Off. Publ. Am. Occup. Ther. Assoc. 51, 530–537. Laing, D.G., 1983. Natural sniffing gives optimum odour perception for humans. Perception 12, 99–117. Martineau, J., Hernandez, N., Hiebel, L., Roché, L., Metzger, A., Bonnet-Brilhault, F., 2011. Can pupil size and pupil responses during visual scanning contribute to the diagnosis of autism spectrum disorder in children? J. Psychiatr. Res. 45, 1077–1082. McCaffrey, R.J., Duff, K., Solomon, G.S., 2000. Olfactory dysfunction discriminates probable Alzheimer's dementia from major depression: a cross-validation and extension. J. Neuropsychiatry Clin. Neurosci. 12, 29–33. McLean, S., 1987. Assessing Dementia. Part I. Aust. N. Z. J. Psychiatry 21 (2), 142–174. Mesholam, R.I., Moberg, P.J., Mahr, R.N., Doty, R.L., 1998. Olfaction in neurodegenerative disease: a meta-analysis of olfactory functioning in Alzheimer's and Parkinson's diseases. Arch. Neurol. 55, 84–90. Moberg, P.J., Doty, R.L., Turetsky, B.I., Arnold, S.E., Mahr, R.N., Gur, R.C., Bilker, W., Gur, R.E., 1997. Olfactory identification deficits in schizophrenia: correlation with duration of illness. Am. J. Psychiatry 154, 1016–1018. Møller, P., Dijksterhuis, G., 2003. Differential human electrodermal responses to odours. Neurosci. Lett. 346, 129–132. Motokizawa, F., 1974. Electrophysiological studies of olfactory projection to the mesencephalic reticular formation. Exp. Neurol. 44, 135–144. Murphy, P.R., O'Connell, R.G., O'Sullivan, M., Robertson, I.H., Balsters, J.H., 2014. Pupil diameter covaries with BOLD activity in human locus coeruleus. Hum. Brain Mapp. 35 (8), 4140–4154. Naudin, M., Mondon, K., El-Hage, W., Desmidt, T., Jaafari, N., Belzung, C., Atanasova, B., 2014. Long-term odor recognition memory in unipolar major depression and Alzheimer s disease. Psychiatry Res. 220 (3), 861–866.

9

Pause, B.M., Miranda, A., Göder, R., Aldenhoff, J.B., Ferstl, R., 2001. Reduced olfactory performance in patients with major depression. J. Psychiatr. Res. 35, 271–277. Pentzek, M., Grass-Kapanke, B., Ihl, R., 2007. Odor identification in Alzheimer's disease and depression. Aging Clin. Exp. Res. 19, 255–258. Price, J.L., Davis, P.B., Morris, J.C., White, D.L., 1991. The distribution of tangles, plaques and related immunohistochemical markers in healthy aging and Alzheimer's disease. Neurobiol. Aging 12, 295–312. Rogers, S.J., Hepburn, S., Wehner, E., 2003. Parent reports of sensory symptoms in toddlers with autism and those with other developmental disorders. J. Autism Dev. Disord. 33, 631–642. Salvucci, D.D., Goldberg, J.H., 2000. Identifying fixations and saccades in eye-tracking protocols. Proceedings of the 2000 Symposium on Eye Tracking Research & Applications. ACM, pp. 71–78. Schneider, C.B., Ziemssen, T., Schuster, B., Seo, H.-S., Haehner, A., Hummel, T., 2009. Pupillary responses to intranasal trigeminal and olfactory stimulation. J. Neural Transm. 116, 885–889 (Vienna Austria 1996). Seigneuric, A., Durand, K., Jiang, T., Baudouin, J.-Y., Schaal, B., 2010. The nose tells it to the eyes: crossmodal associations between olfaction and vision. Perception 39, 1541–1554. Seo, H.-S., Roidl, E., Müller, F., Negoias, S., 2010. Odors enhance visual attention to congruent objects. Appetite 54, 544–549. Shock, N.W., Coombs, C.H., 1973. Changes in skin resistance and affective tone. Am. J. Psychol. 611–620. Solomon, G.S., Petrie, W.M., Hart, J.R., Brackin, H.B., 1998. Olfactory dysfunction discriminates Alzheimer's dementia from major depression. J. Neuropsychiatry Clin. Neurosci. 10, 64–67. Steinhauer, S.R., Siegle, G.J., Condray, R., Pless, M., 2004. Sympathetic and parasympathetic innervation of pupillary dilation during sustained processing. Int. J. Psychophysiol. 52, 77–86. Ward, C.D., Hess, W.A., Calne, D.B., 1983. Olfactory impairment in Parkinson's disease. Neurology 33, 943–946. Warner, M.D., Peabody, C.A., Flattery, J.J., Tinklenberg, J.R., 1986. Olfactory deficits and Alzheimer's disease. Biol. Psychiatry 21, 116–118. Winneke, G., 1992. Structure and determinants of psychophysiological response to odorant/irritant air pollution. Ann. N. Y. Acad. Sci. 641, 261–276. World, M.A.G.A., 2004. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. J. Int. Bioéthique Int. J. Bioeth. 15, 124.

Please cite this article as: Aguillon-Hernandez, N., et al., An odor identification approach based on event-related pupil dilation and gaze focus, Int. J. Psychophysiol. (2015), http://dx.doi.org/10.1016/j.ijpsycho.2015.03.009