Probing the influence of unconscious fear-conditioned visual stimuli on eye movements

Consciousness and Cognition 46 (2016) 60–70

Contents lists available at ScienceDirect

Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog

Probing the influence of unconscious fear-conditioned visual stimuli on eye movements Apoorva Rajiv Madipakkam a,b,⇑, Marcus Rothkirch a, Gregor Wilbertz a, Philipp Sterzer a,c a

Visual Perception Laboratory, Department of Psychiatry, Charité – Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany International Graduate Program Medical Neurosciences, Charité – Universitätsmedizin Berlin, Luisenstr. 56, 10117 Berlin, Germany c Bernstein Center for Computational Neuroscience, Philippstr. 13, 10115 Berlin, Germany b

a r t i c l e

i n f o

Article history: Received 10 June 2016 Revised 8 September 2016 Accepted 17 September 2016

Keywords: Unconscious processing Saccades Face perception Fear conditioning Skin conductance responses Continuous flash suppression

a b s t r a c t Efficient threat detection from the environment is critical for survival. Accordingly, fearconditioned stimuli receive prioritized processing and capture overt and covert attention. However, it is unknown whether eye movements are influenced by unconscious fearconditioned stimuli. We performed a classical fear-conditioning procedure and subsequently recorded participants’ eye movements while they were exposed to fearconditioned stimuli that were rendered invisible using interocular suppression. Chancelevel performance in a forced-choice-task demonstrated unawareness of the stimuli. Differential skin conductance responses and a change in participants’ fearfulness ratings of the stimuli indicated the effectiveness of conditioning. However, eye movements were not biased towards the fear-conditioned stimulus. Preliminary evidence suggests a relation between the strength of conditioning and the saccadic bias to the fear-conditioned stimulus. Our findings provide no strong evidence for a saccadic bias towards unconscious fearconditioned stimuli but tentative evidence suggests that such an effect may depend on the strength of the conditioned response. Ó 2016 Elsevier Inc. All rights reserved.

1. Introduction We are constantly exposed to abundant amounts of sensory information. Due to our brain’s limited processing capacity, we need to prioritize information based on its behavioural relevance to be able to take appropriate action. In particular, the efficient detection and evaluation of threatening stimuli, which signal immediate danger in the environment, is critical for survival. There is indeed empirical evidence that threatening stimuli like fearful or angry faces or snakes receive prioritized visual processing and are detected faster than non-threatening stimuli (LoBue, Matthews, Harvey, & Stark, 2014; Öhman, Flykt, & Esteves, 2001; Öhman, Lundqvist, & Esteves, 2001; Yiend, 2010). Besides such inherent threatening aspects of visual stimuli, stimuli can also acquire a threatening value through classical conditioning, that is, by pairing them with an aversive event. Visual stimuli that have previously been associated with an aversive event (conditioned stimuli, CS+), for instance an aversive noise or an electrical shock, can capture and modulate spatial attention (Armony & Dolan, 2002; Koster, Crombez, Van Damme, Verschuere, & De Houwer, 2004) and can be more readily detected (Padmala & Pessoa, 2008) compared to neutral stimuli.

⇑ Corresponding author at: Department of Psychiatry, Charité – Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany. E-mail address: [email protected] (A.R. Madipakkam). http://dx.doi.org/10.1016/j.concog.2016.09.016 1053-8100/Ó 2016 Elsevier Inc. All rights reserved.

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

61

Emotional and fear-conditioned stimuli also influence the oculomotor system. For instance, fixations increase for CS+ stimuli (Hopkins, Schultz, Hannula, & Helmstetter, 2015) and eye movements are made faster to a peripheral fearful picture than a neutral one (Bannerman, Milders, de Gelder, & Sahraie, 2009). Even for threatening stimuli that are not task-relevant, saccade trajectories and latencies are altered by their occurrence (Mulckhuyse, Crombez, & Van der Stigchel, 2013; Mulckhuyse & Dalmaijer, 2015; Schmidt, Belopolsky, & Theeuwes, 2012, 2015b). Together, these results indicate that fear conditioning has the potency to increase the emotional saliency of a visual stimulus and to consequently enhance its effect on the oculomotor system, drawing eye movements towards conditioned stimuli. From an evolutionary perspective, it would be biologically advantageous to rapidly detect threat from the environment in order to activate appropriate behavioural responses within a short period of time. In line with this notion, neurocognitive theories propose that threatening visual stimuli can also be processed without awareness through a subcortical visual pathway to the amygdala (Tamietto & de Gelder, 2010; Tamietto, Pullens, de Gelder, Weiskrantz, & Goebel, 2012). Consistent with this theory, recent research showed faster access to awareness of fear-conditioned stimuli (Gayet, Paffen, Belopolsky, Theeuwes, & Van der Stigchel, 2016). While such preferential access to awareness does not unequivocally imply enhanced unconscious processing (Gayet, Van der Stigchel, & Paffen, 2014; Sterzer, Stein, Ludwig, Rothkirch, & Hesselmann, 2014), there is also evidence for autonomic responses to fear-conditioned stimuli that were permanently suppressed from awareness (Esteves, Dimberg, & öhman, 1994; Parra, Esteves, Flykt, & Öhman, 1997; Raio, Carmel, Carrasco, & Phelps, 2012; Öhman & Soares, 1993). Furthermore, fear-conditioned visual stimuli, conditioned outside of awareness, have also been shown to modulate visuospatial attention in subsequent tasks (Beaver, Mogg, & Bradley, 2005). Thus, while associations can be formed between aversive unconditioned stimuli (US) and conditioned stimuli (CS) to elicit conditioned responses (CR) outside of awareness, it remains unknown whether fear-conditioned stimuli can modulate overt visual attention and draw eye movements towards them even when presented outside of awareness. Based on results from studies showing physiological responses to fear-conditioned stimuli that were presented outside of awareness (review: Clark, Manns, & Squire, 2002; Critchley, Mathias, & Dolan, 2002) and the oculomotor system’s sensitivity to subliminal stimuli (Rothkirch, Madipakkam, Rehn, & Sterzer, 2015; Rothkirch, Stein, Sekutowicz, & Sterzer, 2012; Spering & Carrasco, 2015), we hypothesized that eye movements would be preferentially directed towards fear-conditioned stimuli that the observers are unaware of. In the current study, we followed a classical fear conditioning procedure (Pavlov, 1927) and manipulated threat by pairing one of two fearful faces (the CS+ stimulus) with an aversive white noise burst (the US stimulus), while never pairing the other face (the CS stimulus) with the noise. Threat was defined in this context as a state of the world predicting an aversive event and the conditioned response as an anticipatory physiological response to the stimulus that predicts an aversive event as learned from prior experience. We chose fearful faces as stimulus material, since the fear conditioning procedure is well established with emotional face stimuli (Critchley et al., 2002; Esteves et al., 1994; Morris, Öhman, & Dolan, 1998; Öhman & Soares, 1993; review: Öhman & Mineka, 2001), and the effectiveness of fear conditioning likely builds upon the ‘preparedness’ of the relation between stimuli such that fear conditioning of threat-related stimuli is acquired faster and is more resistant to extinction than non-threatening stimuli (Öhman, 2009; Seligman, 1971). In addition, fearful faces have elicited reliable SCRs even when conditioned under CFS (Raio et al., 2012). To monitor the effectiveness of the conditioning procedure we assessed physiological responses by recording electrodermal activity during the conditioning phase of the experiment. In a subsequent test phase, we used continuous flash suppression (CFS) (Tsuchiya & Koch, 2005) to render the fear conditioned faces invisible. Eye movements were simultaneously recorded during the test phase to assess whether humans have an oculomotor bias towards unconscious fear-conditioned faces. 2. Material and methods 2.1. Participants Twenty-nine participants took part in the study. The data from two participants were excluded due to poor eye tracking quality. Data from another three participants were excluded due to insufficient suppression of the stimulus (see Section 2.4.2). Thus, the final sample consisted of twenty-four participants (15 female; mean age: 26.41 (±0.99 SEM) years). This sample size was chosen based on a power analysis with the software Gpower V3.1.9.2 (Faul, Erdfelder, Lang, & Buchner, 2007) using an effect size of 0.6 as determined from previous studies that investigated eye movements to stimuli under interocular suppression (Rothkirch et al., 2012, 2015), a power of 0.8 and an alpha level of 0.05. All participants had normal or corrected-to-normal vision, were naïve to the purpose of the study, and were paid for their participation. The study was approved by the local ethics committee and written informed consent was obtained from all participants before the experiment. 2.2. Stimuli The stimuli in the main experimental session consisted of four fearful female faces taken from the Nimstim dataset (Tottenham et al., 2009). The images were converted to grayscale and cropped into oval shapes comprising a size of 3.5°  4.5°. Two versions of each face were created, which only differed in their luminance contrast. The low-level image

62

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

properties were matched across all images of a particular contrast based on their luminance histograms using the SHINE toolbox (Willenbockel et al., 2010). The high contrast face exemplars (root mean square (RMS) contrast of 0.15) were used in the conditioning runs, whereas the low contrast face stimuli (RMS contrast of 0.01) were used in the test runs (see Section 2.3). The visual masks used for CFS in the test runs (see Section 2.3) consisted of randomly distributed black, white, and grey circles (sizes ranging from 0.01° to 0.97°) constituting a square of 12.5°  12.5° (Michelson contrast: 1) and were flashed at a frequency of 10 Hz. For the conditioning runs, the unconditioned stimulus was a white noise burst that had a mean loudness of 70 (±0.80 SEM) dB. Before the start of the first conditioning run, the loudness of the white noise burst was set for each participant to a level that was ‘‘uncomfortable and aversive but not painful”. This resulted in a mean loudness of 70 dB across all participants with a maximum loudness of 75 dB and a minimum loudness of 65 dB. All visual stimuli were presented with Matlab (The MathWorks, Natick, MA, USA), using the Cogent 2000 toolbox (www. vislab.ucl.ak.uk/cogent.php) and displayed on a 19-in. CRT monitor (resolution: 1024  768 Px; refresh rate: 60 Hz). 2.3. Design and procedure The experimental session consisted of 10 runs; 2 preconditioning ‘baseline’ runs, followed by alternating conditioning and test runs, four of each. For each participant, two of the four faces were used, one (CS+) that was paired with an aversive noise in the conditioning runs and the other (CS ) that was never paired with an aversive noise. Participants viewed the computer screen located 50 cm away through a mirror stereoscope, which provided separate input to the two eyes. Participant’s head was stabilized by a chin rest and their eye movements during the preconditioning and test runs were recorded with a high-speed video-based eye tracker (Cambridge Research Systems, UK; sampling rate: 250 Hz; spatial accuracy: 0.05°). Before the start of the main experiment, each participant’s dominant eye was determined in an eye dominance test (Yang, Blake, & McDonald, 2010) using CFS (Tsuchiya & Koch, 2005). The technique uses high contrast dynamic images flashed to one eye to suppress a static stimulus presented to the other eye from awareness through interocular suppression. On each trial of the eye dominance test, dynamic high-contrast greyscale mask stimuli that filled a white square (5°  5°) were flashed to one eye at a frequency of 10 Hz, while simultaneously a face was presented in one quadrant of the square to the other eye. None of the face stimuli from the main experiment were used in the eye dominance test. Participants had to react as fast and accurately as possible to the location of the face stimulus as soon as it overcame suppression and became visible. The eye to which the face stimulus was presented was randomised and counterbalanced across trials. The eye corresponding to shorter reaction times while viewing the face stimuli was identified as the participant’s dominant eye. In the main experiment, the same technique of CFS was used to suppress the fearful stimuli (see Section 2.2) from awareness. A representative trial of the pre-conditioning and test run is depicted in Fig. 1B. Each trial in the pre-conditioning and test run started with a 1500 ms white central fixation cross (0.6°  0.6°) inside a white frame (12.5°  12.5°) on a grey background (luminance: 30.06 cd/m2). This was followed by two 800 ms intervals during which high-contrast dynamic greyscale mask stimuli were flashed at a frequency of 10 Hz to the participants’ dominant eye. Simultaneously, in one of the intervals, two low-contrast fearful face stimuli were presented to the non-dominant eye, one in the left half and the other in the right half of the square (eccentricity: 3.6°). The interval without the face stimuli consisted of the presentation of the white square with a grey background to the non-dominant eye. Both faces were presented at the same vertical position, that is, either 3.7° above, below, or at the horizontal meridian. Participants were instructed to actively search for the faces by making eye movements. Both 800 ms intervals were followed by the presentation of masks to both eyes for 200 ms to reduce afterimages. The two intervals were separated by a fixation period of 750 ms. After the intervals, two questions followed. In the first, participants had to indicate in a manual 2-alternative forced-choice (2AFC) task which of the two intervals contained the faces. This was followed by a rating of their confidence on a four-point confidence scale regarding their 2AFC response, which provided a graded and reliable measure of subjective awareness (Szczepanowski, Traczyk, Wierzchon´, & Cleeremans, 2013). Participants performed 48 trials in each of the pre-conditioning and test runs. Prior to the start of each run, a nine-point calibration of the eye tracker was conducted. The spatial location of the faces as well as their allocation to one of the two intervals was randomised and counterbalanced within each run. A conditioning run (Fig. 1A) always preceded a test run. In the four conditioning runs, each trial started with the 500 ms presentation of a white square (12.5°  12.5°) on a grey background (luminance: 30.06 cd/m2) and a white fixation cross in the centre. Delayed conditioning was employed where one of the fearful faces (CS+) terminated with an aversive noise burst (US) on 50% of the trials (paired CS+ trials) while the other stimulus was never paired with the noise (CS ). With this design of partial reinforcement, providing paired and unpaired CS+ trials, effects of conditioning could be assessed based on unpaired trials only, without the contamination of physiological responses to the aversive noise. The same face pair that was used in the pre-conditioning runs was also used in the conditioning runs. The allocation of a face as CS+ and CS was randomised and counterbalanced across participants. Following the 500 ms pre-stimulus fixation, the face stimulus was presented at high contrast (RMS contrast of 0.15) for 2000 ms. Trials were separated by an 8000 ms inter trial interval. The paired CS+ trials terminated with the presentation of an aversive noise burst which had a mean loudness of 70 dB and which was presented for 250 ms. The order of stimuli was pseudo-randomised such that the same trial type never occurred more than three times consecutively. The first conditioning run had a total of 32 trials (8 paired CS+, 8 unpaired CS+, 16 CS ). Subsequent conditioning runs, which were intended as ‘‘refresher runs”, had half the number of trials, i.e. 16 trials in total (4 paired CS+, 4 unpaired CS+, 8 CS ). During all the conditioning runs, skin conductance responses were recorded with the

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

63

Fig. 1. Trial structure. (A) A typical sequence of stimulus presentation of two consecutive trials in the conditioning phase, one in which a CS+ stimulus was presented followed by a trial in which the CS stimulus was presented. CS+ refers to the conditioned stimulus that was paired on 50% of trials with the unconditioned stimulus (US), the white noise burst. CS refers to the conditioned stimulus that was never paired with the US. (B) A representative trial in the test phase. The face stimuli are depicted at higher contrast and are enlarged for illustration purposes.

Vision Recorder software (Brain Products GmbH, Munich, Germany; sampling rate: 500 Hz) on a mobile notebook. Two Ag-AgCl electrodes were filled with standard NaCl electrolyte paste and placed on the distal phalanges of the second and third fingers of the left hand. Before the start of the first preconditioning run and upon completion of the last test run, participants were asked to rate the intensity of the fearful expression of each face using a visual analogue scale (with anchors ‘not intense’ and ‘very intense’) as a subjective measure for the effectiveness of conditioning. 2.4. Data analyses 2.4.1. Analysis of behavioural data For all analyses, outliers were defined using the interquartile range. Thus, observations that were 1.5 times the interquartile distance above the 75th or below the 25th percentile were considered outliers. When outliers were present, this has been mentioned explicitly in the results. The change in the ratings of the intensity of the fearful expression was computed individually for the CS+ and CS face with the following equation: 100 ⁄ (rating after conditioning rating before condition)/(rating before conditioning). This index indicated the change in the subjective affective value of the CS+ and CS stimuli. A positive index indicates that the perceived intensity of the fearful expression increased after conditioning. For two of the final sample of twenty-four participants the post-conditioning ratings did not get saved due to the computer crashing resulting in an index that was computed for twenty-two participants. Participants’ awareness of the face stimuli was assessed based on both subjective and objective criteria, since an assessment based solely on subjective criteria is prone to response biases (Kunimoto, Miller, & Pashler, 2001; Sterzer et al., 2014). To this end, participants’ performance in the 2AFC task in trials where they indicated to be least confident was evaluated. Participants with an insufficient number of trials in which they indicated to be least confident were excluded from all further analyses. This was determined based on the outlier criteria defined above. Additionally, when performance in the 2AFC task was significantly above chance level based on a binomial test against 50%, participants could no longer be considered objectively unaware of the stimuli and were also excluded from all subsequent analyses. Based on these subjective and objective criteria, three participants had to be excluded. We further tested whether the group mean of the included participants was at chance level. Since a non-significant t-test cannot provide unequivocal evidence in favour of the null hypothesis, we additionally performed a Bayesian analysis (Dienes, 2011; Sterzer et al., 2014) using the statistical software JASP 0.7.5.5. Bayes

64

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

factors (BF10) <0.33 provide substantial evidence for the null over the alternative hypothesis, BF10 >3 can be interpreted as evidence for the alternative over the null hypothesis (Dienes, 2011). 2.4.2. Analysis of eye tracking data Pre-processing of eye tracking data comprised interpolation of missing eye tracking data points if no more than 24 ms of consecutive data were missing using a cubic spline interpolation and then low-pass filtering using a second order SavitzkyGolay filter (Holmqvist et al., 2011). Trials were included in the analysis if (i) Participants made a manual response in both the 2AFC task and the confidence rating; (ii) Participants indicated the lowest confidence rating; and (iii) at least 98% of the eye tracking data collected during a trial after pre-processing were available and not lost due to artefacts. For the included trials, the number of first saccades directed towards the CS+ and CS face before and after conditioning were then calculated. We focused on the first saccade within each trial as it has been demonstrated in previous studies that emotionally and motivationally salient stimuli elicit effects specifically on initial shifts of overt attention (Mulckhuyse & Dalmaijer, 2015; Rothkirch et al., 2015). An eye movement was defined as a saccade when at least 4 consecutive data points (i.e. 16 ms) exceeded a velocity of 40°/s and the euclidian distance between the start and end of the data segment exceeded 0.75°. Anticipatory saccades were defined as eye movements starting earlier than 80 ms after stimulus onset and were excluded from the analysis. To quantify a saccadic bias towards one of the two faces after conditioning, we calculated a normalised saccadic index for the CS+ and CS faces separately based on the proportion of first saccades that landed on the two faces before and after conditioning. The index was defined as 100 ⁄ (sa sb)/(sa + sb) where sa is the proportion of first saccades that landed on the face after conditioning and sb is the proportion of first saccades that landed on the face before conditioning. Positive indices signify an increase in saccades towards the faces after conditioning. However, a significantly larger positive index for CS+ compared to CS would indicate an attentional bias towards the CS+ stimulus after conditioning. Additionally, to avoid insensitivity to eye movement biases that are not related to the first saccade we also determined the average dwell time on the stimulus. The dwell time was defined as the mean percentage of gaze positions directed towards the CS+ and CS . 2.4.3. Analysis of skin conductance data Skin conductance (SC) data acquired from the vision recorder were analysed using Ledalab V3.4.8 (http://www.ledalab. de/), a Matlab based software. The raw SC data were first down-sampled to 10 Hz and low pass filtered (first order Butterworth filter, cut-off frequency 25 Hz). A continuous decomposition analysis (CDA) was performed to separate the SC data into its tonic and phasic components (Benedek & Kaernbach, 2010). Skin conductance response (SCR) peaks within a response window of 1–5.5 s after stimulus onset and with a minimum amplitude threshold 0f 0.03 lS were included in the analysis. In order to reduce inter-individual differences, SCRs were range-corrected for each individual by dividing the individual SCR responses by the largest response amplitude recorded during the experiment, which was usually that of the first US (Lykken & Venables, 1971). Range-corrected SCR amplitudes were square-root transformed to reduce skewness. In order to assess conditioning, a normalised SCR index based on the SCR responses during the unpaired CS+ trials and the CS trials was computed. This index was defined as ([unpaired CS+ minus CS ]/[unpaired CS+ plus CS ]) in line with previous work (Raio et al., 2012). 3. Results 3.1. Effectiveness of conditioning In order to evaluate the effectiveness of our fear-conditioning procedure, we first assessed the physiological and behavioural effects of conditioning by analysing the SCRs elicited by the stimuli and the ratings of the perceived fearfulness of the faces before and after conditioning, respectively. The latter indicates to what degree the subjectively experienced affect of the stimuli was changed by the conditioning procedure. 3.1.1. SCR data SCRs to unpaired CS+ trials (i.e. that did not terminate with the aversive noise, M = 0.33 lS ± 0.01 SEM) were significantly larger than responses to CS trials (M = 0.31 lS ± 0.01 SEM; t[23] = 2.3, p = 0.03; Fig. 2A). This indicates that the presentation of the face that was paired with an aversive noise led to a higher physiological arousal than the presentation of the face that was never paired with the noise. The observed mean SCRs are within the range of previously reported range-corrected SCRs. For example, Öhman & Soares reported average SCRs in the range of 0.22–0.57 to snakes and spider stimuli (Öhman & Soares, 1998). Raio and colleagues reported SCRS to fearful faces in the range of 0.13–0.43 across the different conditions of the experiment (Raio et al., 2012). 3.1.2. Change in fearfulness ratings While the perceived intensity of the fearful expression increased by 9.8% ± 2.9 SEM for the CS+ stimulus, the perceived intensity for the CS stimulus decreased by 9.7% ± 2.9 SEM. This difference between the ratings to the CS+ and CS stimuli was statistically significant (t[21] = 3.32, p = 0.003; Fig. 2B).

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

65

Fig. 2. Responses to the fear-conditioned stimuli. (A) SCRs elicited by the CS+ and CS stimuli. (B) Significant difference in the change in participants’ subjective rating of the fearfulness of the face stimuli before and after conditioning. (C) Saccadic preference indices for the CS+ and CS face. Positive indices indicate an increase in number of saccades to the face stimuli after conditioning. Asterisks (*) indicate statistically significant differences: *p < 0.05; ** p < 0.001. Error bars denote within-subject SEM (Cousineau, 2005).

These results indicate that the fear conditioning procedure was effective in eliciting differential SCRs and changing the subjectively experienced affect of the CS+ and CS stimulus. 3.2. Responses to stimuli presented under CFS 3.2.1. Behavioural data To ensure participants’ unawareness of the face stimuli during the test runs, analyses focused only on those trials where they indicated to be least confident (which was on average, in 81.2% ± 2.4 SEM of all trials). In these ‘very unsure’ trials, participants were not able to correctly guess the presentation of the faces in the manual 2AFC task with above-chance accuracy (M = 49.52% ± 0.74 SEM; one-sample t-test against the chance level of 50%: t[23] = 0.63, p = 0.53), indicating that the faces (in these trials) were effectively suppressed from awareness. Additionally, we performed a Bayesian analysis to test that the data were neither significantly above or below chance. A Bayes factor (BF10) of 0.25 (i.e. <0.33) provided substantial evidence for the null hypothesis, that is, no significant difference from chance level. 3.2.2. Eye tracking data For the analysis of eye tracking data, we again focused on the trials in which participants were unaware of the presentation of the faces, that is, trials in which they expressed to be least confident according to their rating and also not above chance in their 2AFC discrimination response (see Section 3.2.1). We first probed whether eye movements were influenced by the general presence of the face stimuli by comparing the number of first saccades that were made in the interval with and without the stimuli. The proportion of trials in which at least one saccade was made in the interval with the stimuli (M = 75.16% ± 0.89 SEM) was significantly higher than in the interval without stimuli (M = 65.92 ± 0.89 SEM; t[23] = 5.21, p < 0.001), suggesting that participants’ eye movement behaviour was affected by the presence of the face stimuli. Next we investigated whether the saccades made in the interval with the face stimuli were more frequently directed to the actual positions of the faces in comparison to all possible locations where the two faces could appear. To this end, we computed the proportion of first saccades that landed on the two concurrently presented faces in relation to the proportion of saccades that landed within all possible face locations. Saccades showed a tendency towards being directed more frequently to the locations in which the faces were actually presented, compared to all possible locations of the faces (M = 38.8% ± 1.5 SEM; one sample t-test against 36% representing the area covered by the faces with respect to all possible locations: t[23] = 1.87, p = 0.07). To test the specific effects of fear conditioning on saccadic eye movements, we computed individual saccadic indices for the CS+ and CS face. The index was defined as the ratio of the difference between the proportion of saccades directed to the face stimulus after and before conditioning to the sum of the saccadic proportions after and before conditioning. Thus, positive indices indicate an increase in saccades to the stimulus after conditioning. The change in the relative frequency of saccades directed towards the CS+ face (M = 23.2% ± 6.3 SEM) was not statistically different from the change in the frequency of saccades directed towards the CS face (M = 19.9% ± 6.3 SEM; t[23] = 0.27, p = 0.80; Fig. 2C). The lack of a significant difference between saccades towards CS+ and CS faces was further corroborated by a Bayes factor (BF10) of 0.22, indicating substantial evidence for the null hypothesis. It is interesting to note, however, that the saccadic indices of both the CS+ and CS face were positive indicating a general increase in saccades towards the faces after conditioning. Only

66

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

Fig. 3. Correlation between the difference in the saccadic index to the CS+ and CS stimulus and the normalised SCR index. A positive correlation suggests that the saccadic bias towards the CS+ face after conditioning depends on the strength of the conditioned response. Unfilled circles represent outliers along the x dimension. Inclusion of the outliers results in a positive correlation that is no longer significant (r = 0.21, p = 0.34).

the saccadic index for the CS+ face was significantly above 0 (t[23] = 2.5, pcorr = 0.04, one sample t-test against 0, which would have indicated no difference in the frequency of saccades between before and after conditioning). Although the index for the CS face was also positive, it did not survive Bonferroni correction for multiple comparisons. (t[23] = 2.1, pcorr = 0.09; Note: pcorr refers to the Bonferroni corrected p value, corrected for the number of tests, in this case, two). Further, there was no difference in the change in the dwell times (i.e. the mean percentage of gaze positions directed towards the stimulus) between the CS+ (M = 0.13% ± 0.55 SEM) and CS (M = 0.39% ± 0.55 SEM) stimulus after conditioning (t[23] = 0.38, p > 0.70; BF10 = 0.22). Thus, although there was a general saccadic bias in the interval with stimuli with significantly more saccades made and a tendency for these saccades to be directed to the faces more often than the other face locations, there was no specific saccadic bias with respect to the fear-conditioned face. We also investigated whether a saccadic bias to the CS+ stimulus might be dependent on the strength of conditioning. To this end, we correlated the difference in the saccadic indices of the CS+ and CS stimuli to the difference in the SCR indices of the stimuli. The saccadic index represents the difference in the proportions of saccades that were directed to the CS+ and CS face before and after conditioning. To rule out that the strength of the conditioned response (i.e. the SCR) could be driven by participants’ awareness of the stimulus (cf. Cornwell, Echiverri, & Grillon, 2007; Knight, Nguyen, & Bandettini, 2003), we performed a partial correlation controlling for each participants’ awareness level as indicated by their 2AFC task performance. Based on the outlier criteria defined in Section 2.4.1, two outliers were revealed in the saccadic indices (Fig. 3, unfilled circles). Thus, these data points were excluded from the correlation. We observed a significant positive correlation, indicating that the attentional bias to the CS+ face after conditioning depended on the strength of the conditioning procedure as measure by the skin conductance response (Pearson’s r = 0.48, p = 0.02) (Fig. 3). However, we also repeated the correlation including the two outliers, which was - although still positive - not statistically significant anymore (Pearson’s r = 0.21, p = 0.33). Thus, this result needs to be interpreted with caution. There was no correlation between the saccadic indices and the participants’ change in the subjectively experienced affect of the stimuli (r = 0.06, p > 0.5). Finally, we repeated the analysis of eye movements towards conditioned stimuli only for the subsample of participants who had a positive SCR index, i.e., a higher SCR to the CS+ stimulus than to the CS stimulus. The rationale for this analysis followed from the correlation results that suggested that an attentional bias to the CS+ stimulus might be dependent on the strength of the conditioned response. Of the total sample of 24 participants, 17 had a positive SCR index. In this subsample, the relative frequency of saccades directed towards the CS+ face (M = 29.4% ± 8.3 SEM) was again not statistically different from the frequency of saccades directed to the CS face (M = 17.8% ± 8.3 SEM; t[16] = 0.70, p = 0.49). The Bayes factor for this analysis was 0.30 indicating evidence for the null hypothesis. These results suggest that the lack of a saccadic bias to the CS+ stimulus is not only due to those participants who did not have a positive conditioned response. While a positive SCR index might be necessary to cause a saccadic bias to a suppressed conditioned stimulus, it does not seem to be sufficient. Rather, a particular threshold or a sufficiently strong conditioning effect might need to be reached. 4. Discussion In the current study, we investigated whether fear-conditioned face stimuli influence eye movements even when presented outside of awareness. We found significant effects of fear conditioning on skin conductance responses and subjective fearfulness ratings, showing that our conditioning procedure was effective. When assessing the effect of fear conditioning on saccades during suppression of the face stimuli from awareness, we found a general increase in saccades to both the CS+ and CS stimuli after conditioning. However, we observed no specific saccadic bias towards the CS+ face stimuli after conditioning. Preliminary evidence showed a relationship between the strength of conditioning as measured by the skin conductance response and eye movements towards the CS+ stimulus; larger conditioned responses were associated with a positive saccadic index indicating a bias towards the fear-conditioned stimulus.

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

67

Previous studies using the classical fear conditioning paradigm for fully visible stimuli have shown that fear-conditioned stimuli capture and hold covert attention (Koster et al., 2004; Schmidt, Belopolsky, & Theeuwes, 2015a; Van Damme, Crombez, & Notebaert, 2008). Furthermore fear-conditioned stimuli also influence overt attentional selection. For example, fixations increase for CS+ stimuli (Hopkins et al., 2015) and additionally, these stimuli influence attention and visual selection by altering saccade trajectories (Mulckhuyse & Dalmaijer, 2015; Mulckhuyse et al., 2013) indicating that such fear-conditioned stimuli have an effect on oculomotor behaviour. Thus, our results suggest that in the specific context of the experiment, i.e. fear-conditioned face stimuli masked with CFS, there was no measurable effect of conditioning on eye movements. In the following paragraphs we discuss the potential causes for this absence of an effect at the level of eye movements, beginning with the role of awareness in processing interocularly suppressed stimuli. There is a longstanding (and still ongoing) debate about the extent to which information can be processed outside of awareness and can influence behavioural responses (Hesselmann & Moors, 2015; Lovibond & Shanks, 2002, Pessoa, 2005; Sterzer et al., 2014). In the context of fear conditioning, this discussion has largely been limited to contingency awareness, that is, the observers’ conscious knowledge about the association between the conditioned and the unconditioned stimulus (Lovibond & Shanks, 2002; Purkis & Lipp, 2001). In the current study, we did not directly assess contingency awareness as it was not relevant to our research question. The critical question in our study was to which extent aversive information acquired by fear-conditioning can influence behaviour in the complete absence of awareness of the stimuli. Previous studies that used backward masking to suppress visual stimuli from awareness showed enhanced physiological responses, like SCRs and startle reflex, to ‘unconscious’ fear-conditioned stimuli (Esteves et al., 1994; Olsson & Phelps, 2004; Ruiz-Padial & Vila, 2007; Öhman & Soares, 1998). However, these studies have been criticized for not adequately assessing perceptual awareness (Grillon & Cornwell, 2007; Lovibond & Shanks, 2002; Pessoa, 2005). The above mentioned studies assumed that backward masking with target presentation times of 17–33 ms would reliably suppress stimuli from awareness. However, signal detection tasks show variable effectiveness of backward masking of emotional faces where masked targets can be discriminated even at extremely short stimulus onset asynchronies (Maxwell & Davidson, 2004; Pessoa, Japee, & Ungerleider, 2005) and thus cannot be assumed to be fully invisible. Furthermore, SCRs to CS+ stimuli depend on the strength of suppression of the stimulus, with higher conditioned responses to stimuli that are less suppressed (Cornwell et al., 2007). Specifically with respect to the masking technique CFS, Raio and colleagues observed differential SCRs to fearful faces that were conditioned outside of awareness by rendering them invisible using CFS (Raio et al., 2012). Furthermore, the authors assessed participants’ awareness based on both subjective and objective measures similar to the current study. However, the nature of the 2AFC task used in their study to assess objective awareness does not ultimately rule out partial awareness of the stimuli. More precisely, participants were asked to report the gender of the suppressed faces. Critically though, chance performance in a gender categorization task, does not necessarily preclude partial perception of the stimulus. Thus, low-level awareness of certain stimulus features could still drive differential SCRs independently of forming a representation necessary for gender identification (Lovibond & Shanks, 2002). In a similar vein, Gelbard-Sagiv, Faivre, Mudrik, and Koch (2016) claim that high-level processing under CFS is only enabled by awareness of low-level stimulus features (Gelbard-Sagiv et al., 2016). In other words, the unawareness of high-level processes under CFS does not directly imply low-level unawareness. In this context, the absence or presence of a saccadic bias in trials in which participants reported at least some degree of stimulus awareness in our study could have been informative. However, our experiment was designed to maximise unawareness of the face stimuli. As a consequence, participants reported awareness of the faces only in a small proportion of trials, which precludes a meaningful analysis. To specifically address the question to what extent eye movements towards fearconditioned stimuli are dependent on the degree of visibility of these stimuli, a more suitable approach would be to adjust the mask contrast to obtain similar numbers of visible and invisible trials (Madipakkam, Rothkirch, Guggenmos, Heinz, & Sterzer, 2015). Thus far, the results from previous studies seem to suggest that conditioned responses to suppressed stimuli might be modulated by awareness. However, Raio and colleagues as well as previous backward masking studies have provided evidence of differential SCRs to suppressed stimuli. Together with the results from the current study, this leaves open the question whether fear-conditioned stimuli presented outside awareness only have an influence at the physiological level or whether other more behaviour-relevant functions such as eye movements are also influenced. In our present study, larger SCRs were observed for the CS+ in comparison to the CS stimulus during the conditioning phase, where the stimuli were visible, showing that the conditioning procedure was effective. Although eye movements were not preferentially directed to the CS+ stimulus when presented outside of awareness, it might have been possible to detect differential SCRs to the CS+ and CS stimuli also during the test phase. However, the simultaneous presentation of the CS+ and CS stimuli during the test phase, precluded the differential assessment of physiological responses during this phase of the experiment. On the other hand, the simultaneous presentation of the stimuli allowed a direct comparison of eye movements within the same trial. An experimental design with the test phase having three types of trials under CFS (only CS +, only CS and a simultaneous presentation of both CS+ and CS ) might have enabled the measurement of SCRs under CFS together with an intra-trial comparison of eye movements. However, our experiment was designed to minimize the known decrease in the effectiveness of CFS (Ludwig, Sterzer, Kathmann, Franz, & Hesselmann, 2013; Mastropasqua, Tse, & Turatto, 2015) but such a design prevented us from assessing SCR effects simultaneously with eye movement effects. It is also possible that fear conditioning during the acquisition phase did not transfer to the test phase due to perceptual differences in stimuli between the two phases. Alternatively, conditioned responses to suppressed stimuli could only have an influence at the physiological level as shown by Raio et al. (2012) and do not influence eye movements.

68

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

Yet another reason for the observed null effect could be the choice of stimuli used in the current experiment. We used fearful faces in the experiment as such emotionally negative stimuli have been shown to produce reliable conditioned responses to the fear conditioning procedure (Esteves et al., 1994; Öhman & Soares, 1993; review: Öhman & Mineka, 2001). Importantly, negative stimuli are more resistant to extinction, which was particularly relevant in the current experimental design as no unconditioned stimulus was delivered during the test phase. Thus, a saccadic bias towards CS+ stimuli would largely rest upon a carry-over effect of fear conditioning to the test phase, which means that it should to some extent, be resistant to extinction. Importantly, Raio et al. (2012), used the same suppression technique of CFS to show that fearful stimuli can elicit differential SCRs even when the conditioning procedure is performed under suppression. While the use of fearful faces may have driven conditioning, it is plausible that eye movements could have been attracted to these faces independent of conditioning. Indeed, previous research has shown that eye movements are in general attracted to stimuli suppressed from awareness, even to low-level stimuli such as gabor patches (Rothkirch et al., 2012). In line with this, two findings in our experiment suggest that the stimuli were processed nonspecifically under CFS. First, more saccades were performed in the interval in which the stimuli were presented compared to the suppressed blank interval; and second, saccades tended to be directed to the face stimuli more often than to other locations where the faces could have been presented. Moreover, even if eye movements are attracted to neutral stimuli, they may be even more strongly attracted to fearful face stimuli, resulting in a ceiling effect and overshadowing subtle effects of fear conditioning on eye movements. In fact, the recent observation that eye movements are preferentially directed towards suppressed faces with direct gaze indeed suggests that there is room beyond the simple effect of neutral stimuli (Rothkirch et al., 2015). We can speculate that the eyes – as the most salient feature in face stimuli and for fearful faces in particular (due to the enlarged eye whites) – have driven this effect. While information from the eye region can be processed unconsciously (Rothkirch et al., 2015), this information solely would not have been sufficient to cause a saccadic bias to the CS+ face in the current study. To observe a saccadic bias to the CS+ face stimulus after conditioning it would require the processing of face identity under CFS, a more complex process than the processing of a single feature. The extent to which high level information can be processed under CFS is still under discussion. Previous studies using CFS to investigate unconscious face identity processing have shown that awareness might indeed be required for holistic face processing, an essential component of face recognition (Axelrod & Rees, 2014). Other studies have found that face adaptation effects and identity processing also require visual awareness (Izatt, Dubois, Faivre, & Koch, 2014; Moradi, Koch, & Shimojo, 2005; Stein & Sterzer, 2011; review: Axelrod, Bar, & Rees, 2015). Taken together, the results of the current study are in line with previous studies that suggest face identities cannot be processed unconsciously. However, positive saccadic indices to both the CS+ and CS face suggest that even when stimulus information is degraded through CFS, the conditioning procedure causes a general increase in sensitivity to faces. Although eye movements were not influenced by fear-conditioned, high-level visual stimuli like fearful faces when presented outside of awareness, they could still be susceptible to fear-conditioning if the fear-conditioned stimuli were distinguishable based on a simple low-level feature. In line with this idea, there is recent evidence that fear-conditioned grating stimuli access awareness faster than neutral stimuli (Gayet et al., 2016). While this study employed the breaking-CFS paradigm to measure access to awareness which does not unequivocally measure unconscious processing (Gayet et al., 2014; Stein & Sterzer, 2014), it lends support to the idea that low-level fear-conditioned stimuli could also influence the oculomotor system. Thus, it is plausible that an effect of conditioning on eye movements to suppressed stimuli may have been observed with the use of either more dissimilar stimuli and/or low-level stimuli. However, such an interpretation must currently remain speculative. Interestingly, a saccadic bias to the CS+ stimulus was observed in participants in whom conditioning was most effective. This suggests that the influence of fear-conditioned stimuli on eye movements does not depend on awareness but rather on the arousal induced by the stimulus. This finding therefore speaks against the other possible reasons discussed above (awareness, stimuli used, perceptual differences between the conditioning and test phase resulting in a lack of transfer of the conditioned response during the test phase) for the absence of an effect on eye movements. Low arousal as measured by SCRs, may not be sufficient at the level of eye movements, i.e. an attentional biasing towards the CS+ stimulus. Importantly, the relation between the strength of conditioning and eye movements was not influenced by participants’ awareness of the stimulus as observed in previous studies (Cornwell et al., 2007; Knight et al., 2003). However, as these results are only preliminary, future studies need to determine the robustness of this effect. Nevertheless, they could explain the overall absence of a saccadic preference to the CS+ stimulus compared to the CS after conditioning at the group-level. To conclude, the current study investigated whether humans have a saccadic bias to fear-conditioned faces presented outside of awareness. The fear-conditioning procedure was effective in changing the subjective value of the visibly presented face stimuli and in eliciting differential responses to the CS+ and CS stimuli. However, when presented outside of awareness, no specific saccadic bias to the CS+ face after conditioning was observed. There was a general increase in the number of saccades in the interval with stimuli and a tendency for saccades to land more often on the actual faces compared to all possible locations where the faces could be presented indicating the oculomotor systems sensitivity to unaware stimuli. While it is difficult to isolate a single cause for the absence of an effect, our finding of a positive correlation between the SCR indices and the saccadic indices suggests that a saccadic bias towards unconscious fear-conditioned faces may be dependent on the strength of the conditioned response. In other words, we can tentatively conclude that conditioned stimuli may in fact influence eye movements, but only if conditioning is strong enough.

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

69

Funding This work was supported by the German Research Foundation (DFG: STE 1430/7-1 and RO 4836/2-1) and Cluster of Excellence Neurocure (EXC 257). Acknowledgments We thank Kristina Kelly for her help with data acquisition. References Armony, J. L., & Dolan, R. J. (2002). Modulation of spatial attention by fear-conditioned stimuli: An event-related fMRI study. Neuropsychologia, 40(7), 817–826. http://dx.doi.org/10.1016/S0028-3932(01)00178-6. Axelrod, V., Bar, M., & Rees, G. (2015). Exploring the unconscious using faces. Trends in Cognitive Sciences, 19(1), 35–45. http://dx.doi.org/10.1016/j. tics.2014.11.003. Axelrod, V., & Rees, G. (2014). Conscious awareness is required for holistic face processing. Consciousness and Cognition, 27, 233–245. http://dx.doi.org/ 10.1016/j.concog.2014.05.004. Bannerman, R. L., Milders, M., de Gelder, B., & Sahraie, A. (2009). Orienting to threat: Faster localization of fearful facial expressions and body postures revealed by saccadic eye movements. Proceedings of the Royal Society of London B: Biological Sciences, 276(1662), 1635–1641. http://dx.doi.org/10.1098/ rspb.2008.1744. Beaver, J. D., Mogg, K., & Bradley, B. P. (2005). Emotional conditioning to masked stimuli and modulation of visuospatial attention. Emotion, 5(1), 67–79. http://dx.doi.org/10.1037/1528-3542.5.1.67. Benedek, M., & Kaernbach, C. (2010). A continuous measure of phasic electrodermal activity. Journal of Neuroscience Methods, 190(1), 80–91. http://dx.doi. org/10.1016/j.jneumeth.2010.04.028. Clark, R. E., Manns, J. R., & Squire, L. R. (2002). Classical conditioning, awareness, and brain systems. Trends in Cognitive Sciences, 6(12), 524–531. http://dx. doi.org/10.1016/S1364-6613(02)02041-7. Cornwell, B. R., Echiverri, A. M., & Grillon, C. (2007). Sensitivity to masked conditioned stimuli predicts conditioned response magnitude under masked conditions. Psychophysiology, 44(3), 403–406. http://dx.doi.org/10.1111/j.1469-8986.2007.00519.x. Cousineau, D. (2005). Confidence intervals in within-subject designs: A simpler solution to Loftus and Masson’s method. Tutorials in Quantitative Methods for Psychology, 1, 42–45. Critchley, H. D., Mathias, C. J., & Dolan, R. J. (2002). Fear conditioning in humans: The influence of awareness and autonomic arousal on functional neuroanatomy. Neuron, 33(4), 653–663. http://dx.doi.org/10.1016/S0896-6273(02)00588-3. Dienes, Z. (2011). Bayesian versus orthodox statistics: Which side are you on? Perspectives on Psychological Science, 6(3), 274–290. http://dx.doi.org/10.1177/ 1745691611406920. Esteves, F., Dimberg, U., & öhman, A. (1994). Automatically elicited fear: Conditioned skin conductance responses to masked facial expressions. Cognition and Emotion, 8(5), 393–413. http://dx.doi.org/10.1080/02699939408408949. Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G⁄Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. Gayet, S., Paffen, C. L. E., Belopolsky, A. V., Theeuwes, J., & Van der Stigchel, S. (2016). Visual input signaling threat gains preferential access to awareness in a breaking continuous flash suppression paradigm. Cognition, 149, 77–83. http://dx.doi.org/10.1016/j.cognition.2016.01.009. Gayet, S., Van der Stigchel, S., & Paffen, C. L. (2014). Breaking continuous flash suppression: Competing for consciousness on the pre-semantic battlefield. Frontiers in Psychology, 5. Gelbard-Sagiv, H., Faivre, N., Mudrik, L., & Koch, C. (2016). Low-level awareness accompanies ‘‘unconscious” high-level processing during continuous flash suppression. Journal of Vision, 16(1), 3. http://dx.doi.org/10.1167/16.1.3. Grillon, C., & Cornwell, B. (2007). Comments on ‘‘Fearful and sexual pictures not consciously seen modulate the startle reflex in human beings”, by RuizPadial and Vila, online on the Biological Psychiatry web site. Biological Psychiatry, 62(5), 541–542. http://dx.doi.org/10.1016/j.biopsych.2006.12.026. Hesselmann, G., & Moors, P. (2015). Definitely maybe: Can unconscious processes perform the same functions as conscious processes? Frontiers in Psychology, 6. http://dx.doi.org/10.3389/fpsyg.2015.00584. Holmqvist, K., Nyström, M., Andersson, R., Dewhurst, R., Jarodzka, H., & van de Weijer, J. (2011). Eye tracking – A comprehensive guide to methods and measures. Oxford University Press. Hopkins, L. S., Schultz, D. H., Hannula, D. E., & Helmstetter, F. J. (2015). Eye movements index implicit memory expression in fear conditioning. PLoS ONE, 10 (11). http://dx.doi.org/10.1371/journal.pone.0141949. Izatt, G., Dubois, J., Faivre, N., & Koch, C. (2014). A direct comparison of unconscious face processing under masking and interocular suppression. Frontiers in Psychology, 5. http://dx.doi.org/10.3389/fpsyg.2014.00659. Knight, D. C., Nguyen, H. T., & Bandettini, P. A. (2003). Expression of conditional fear with and without awareness. Proceedings of the National Academy of Sciences of the United States of America, 100(25), 15280–15283. http://dx.doi.org/10.1073/pnas.2535780100. Koster, E. H. W., Crombez, G., Van Damme, S., Verschuere, B., & De Houwer, J. (2004). Does imminent threat capture and hold attention? Emotion, 4(3), 312–317. http://dx.doi.org/10.1037/1528-3542.4.3.312. Kunimoto, C., Miller, J., & Pashler, H. (2001). Confidence and accuracy of near-threshold discrimination responses. Consciousness and Cognition, 10(3), 294–340. http://dx.doi.org/10.1006/ccog.2000.0494. LoBue, V., Matthews, K., Harvey, T., & Stark, S. L. (2014). What accounts for the rapid detection of threat? Evidence for an advantage in perceptual and behavioral responding from eye movements. Emotion, 14(4), 816–823. http://dx.doi.org/10.1037/a0035869. Lovibond, P. F., & Shanks, D. R. (2002). The role of awareness in Pavlovian conditioning: Empirical evidence and theoretical implications. Journal of Experimental Psychology: Animal Behavior Processes, 28(1), 3–26. http://dx.doi.org/10.1037/0097-7403.28.1.3. Ludwig, K., Sterzer, P., Kathmann, N., Franz, V. H., & Hesselmann, G. (2013). Learning to detect but not to grasp suppressed visual stimuli. Neuropsychologia, 51(13), 2930–2938. http://dx.doi.org/10.1016/j.neuropsychologia.2013.09.035. Lykken, D. T., & Venables, P. H. (1971). Direct measurement of skin conductance: A proposal for standardization. Psychophysiology, 8(5), 656–672. http://dx. doi.org/10.1111/j.1469-8986.1971.tb00501.x. Madipakkam, A. R., Rothkirch, M., Guggenmos, M., Heinz, A., & Sterzer, P. (2015). Gaze direction modulates the relation between neural responses to faces and visual awareness. The Journal of Neuroscience, 35(39), 13287–13299. http://dx.doi.org/10.1523/JNEUROSCI.0815-15.2015. Mastropasqua, T., Tse, P. U., & Turatto, M. (2015). Learning of monocular information facilitates breakthrough to awareness during interocular suppression. Attention, Perception & Psychophysics, 77(3), 790–803. http://dx.doi.org/10.3758/s13414-015-0839-z. Maxwell, J., & Davidson, R. (2004). Unequally masked: Indexing differences in the perceptual salience of ‘‘unseen” facial expressions. Cognition and Emotion, 18(8), 1009–1026. http://dx.doi.org/10.1080/02699930441000003. Moradi, F., Koch, C., & Shimojo, S. (2005). Face adaptation depends on seeing the face. Neuron, 45(1), 169–175. http://dx.doi.org/10.1016/j. neuron.2004.12.018.

70

A.R. Madipakkam et al. / Consciousness and Cognition 46 (2016) 60–70

Morris, J. s., Öhman, A., & Dolan, R. J. (1998). Conscious and unconscious emotional learning in the human amygdala. Nature, 393(6684), 467. Mulckhuyse, M., Crombez, G., & Van der Stigchel, S. (2013). Conditioned fear modulates visual selection. Emotion, 13(3), 529–536. http://dx.doi.org/10.1037/ a0031076. Mulckhuyse, M., & Dalmaijer, E. S. (2015). Distracted by danger: Temporal and spatial dynamics of visual selection in the presence of threat. Cognitive, Affective, & Behavioral Neuroscience. http://dx.doi.org/10.3758/s13415-015-0391-2. Öhman, A. (2009). Of snakes and faces: An evolutionary perspective on the psychology of fear. Scandinavian Journal of Psychology, 50(6), 543–552. http://dx. doi.org/10.1111/j.1467-9450.2009.00784.x. Öhman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detecting the snake in the grass. Journal of Experimental Psychology: General, 130(3), 466–478. http://dx.doi.org/10.1037/0096-3445.130.3.466. Öhman, A., Lundqvist, D., & Esteves, F. (2001). The face in the crowd revisited: A threat advantage with schematic stimuli. Journal of Personality and Social Psychology, 80(3), 381–396. http://dx.doi.org/10.1037/0022-3514.80.3.381. Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108(3), 483–522. http://dx.doi.org/10.1037/0033-295X.108.3.483. Öhman, A., & Soares, J. J. (1993). On the automatic nature of phobic fear: Conditioned electrodermal responses to masked fear-relevant stimuli. Journal of Abnormal Psychology, 102(1), 121–132. Öhman, A., & Soares, J. J. F. (1998). Emotional conditioning to masked stimuli: Expectancies for aversive outcomes following nonrecognized fear-relevant stimuli. Journal of Experimental Psychology: General, 127(1), 69–82. http://dx.doi.org/10.1037/0096-3445.127.1.69. Olsson, A., & Phelps, E. A. (2004). Learned fear of ‘‘unseen” faces after pavlovian, observational, and instructed fear. Psychological Science, 15(12), 822–828. http://dx.doi.org/10.1111/j.0956-7976.2004.00762.x. Padmala, S., & Pessoa, L. (2008). Affective learning enhances visual detection and responses in primary visual cortex. The Journal of Neuroscience, 28(24), 6202–6210. http://dx.doi.org/10.1523/JNEUROSCI.1233-08.2008. Parra, C., Esteves, F., Flykt, A., & Öhman, A. (1997). Pavlovian conditioning to social stimuli. European Psychologist, 2(2), 106–117. http://dx.doi.org/10.1027/ 1016-9040.2.2.106. Pavlov, I. P. (1927). Conditioned reflexes: An investigation of the physiological activities of the cerebral cortex. New York: Dover. Pessoa, L. (2005). To what extent are emotional visual stimuli processed without attention and awareness? Current Opinion in Neurobiology, 15(2), 188–196. http://dx.doi.org/10.1016/j.conb.2005.03.002. Pessoa, L., Japee, S., & Ungerleider, L. G. (2005). Visual awareness and the detection of fearful faces. Emotion (Washington, D.C.), 5(2), 243–247. http://dx.doi. org/10.1037/1528-3542.5.2.243. Purkis, H. M., & Lipp, O. V. (2001). Does affective learning exist in the absence of contingency awareness? Learning and Motivation, 32(1), 84–99. http://dx. doi.org/10.1006/lmot.2000.1066. Raio, C. M., Carmel, D., Carrasco, M., & Phelps, E. A. (2012). Nonconscious fear is quickly acquired but swiftly forgotten. Current Biology: CB, 22(12), R477–R479. http://dx.doi.org/10.1016/j.cub.2012.04.023. Rothkirch, M., Madipakkam, A. R., Rehn, E., & Sterzer, P. (2015). Making eye contact without awareness. Cognition, 143, 108–114. http://dx.doi.org/10.1016/j. cognition.2015.06.012. Rothkirch, M., Stein, T., Sekutowicz, M., & Sterzer, P. (2012). A direct oculomotor correlate of unconscious visual processing. Current Biology, 22(13), R514–R515. Ruiz-Padial, E., & Vila, J. (2007). Fearful and sexual pictures not consciously seen modulate the startle reflex in human beings. Biological Psychiatry, 61(8), 996–1001. http://dx.doi.org/10.1016/j.biopsych.2006.08.046. Schmidt, L. J., Belopolsky, A. V., & Theeuwes, J. (2012). The presence of threat affects saccade trajectories. Visual Cognition, 20(3), 284–299. http://dx.doi.org/ 10.1080/13506285.2012.658885. Schmidt, L. J., Belopolsky, A. V., & Theeuwes, J. (2015a). Attentional capture by signals of threat. Cognition and Emotion, 29(4), 687–694. http://dx.doi.org/ 10.1080/02699931.2014.924484. Schmidt, L. J., Belopolsky, A. V., & Theeuwes, J. (2015b). Potential threat attracts attention and interferes with voluntary saccades. Emotion, 15(3), 329–338. http://dx.doi.org/10.1037/emo0000041. Seligman, M. E. P. (1971). Phobias and preparedness. Behavior Therapy, 2(3), 307–320. http://dx.doi.org/10.1016/S0005-7894(71)80064-3. Spering, M., & Carrasco, M. (2015). Acting without seeing: Eye movements reveal visual processing without awareness. Trends in Neurosciences, 38(4), 247–258. http://dx.doi.org/10.1016/j.tins.2015.02.002. Stein, T., & Sterzer, P. (2011). High-level face shape adaptation depends on visual awareness: Evidence from continuous flash suppression. Journal of Vision, 11(8), 5. http://dx.doi.org/10.1167/11.8.5. 1–14. Stein, T., & Sterzer, P. (2014). Unconscious processing under interocular suppression: Getting the right measure. Frontiers in Psychology, 5. Sterzer, P., Stein, T., Ludwig, K., Rothkirch, M., & Hesselmann, G. (2014). Neural processing of visual information under interocular suppression: A critical review. Frontiers in Psychology, 5. http://dx.doi.org/10.3389/fpsyg.2014.00453. Szczepanowski, R., Traczyk, J., Wierzchon´, M., & Cleeremans, A. (2013). The perception of visual emotion: Comparing different measures of awareness. Consciousness and Cognition, 22(1), 212–220. http://dx.doi.org/10.1016/j.concog.2012.12.003. Tamietto, M., & de Gelder, B. (2010). Neural bases of the non-conscious perception of emotional signals. Nature Reviews Neuroscience, 11(10), 697–709. http://dx.doi.org/10.1038/nrn2889. Tamietto, M., Pullens, P., de Gelder, B., Weiskrantz, L., & Goebel, R. (2012). Subcortical connections to human amygdala and changes following destruction of the visual cortex. Current Biology, 22(15), 1449–1455. Tottenham, N., Tanaka, J. W., Leon, A. C., McCarry, T., Nurse, M., Hare, T. A., & Nelson, C. (2009). The NimStim set of facial expressions: Judgments from untrained research participants. Psychiatry Research, 168(3), 242–249. http://dx.doi.org/10.1016/j.psychres.2008.05.006. Tsuchiya, N., & Koch, C. (2005). Continuous flash suppression reduces negative afterimages. Nature Neuroscience, 8(8), 1096–1101. http://dx.doi.org/ 10.1038/nn1500. Van Damme, S., Crombez, G., & Notebaert, L. (2008). Attentional bias to threat: A perceptual accuracy approach. Emotion, 8(6), 820–827. http://dx.doi.org/ 10.1037/a0014149. Willenbockel, V., Sadr, J., Fiset, D., Horne, G., Gosselin, F., & Tanaka, J. (2010). Controlling low-level image properties: The SHINE toolbox. Behavior Research Methods, 42(3), 671–684. http://dx.doi.org/10.3758/BRM.42.3.671. Yang, E., Blake, R., & McDonald, J. E. (2010). A new interocular suppression technique for measuring sensory eye dominance. Investigative Opthalmology & Visual Science, 51(1), 588. http://dx.doi.org/10.1167/iovs.08-3076. Yiend, J. (2010). The effects of emotion on attention: A review of attentional processing of emotional information. Cognition and Emotion, 24(1), 3–47. http:// dx.doi.org/10.1080/02699930903205698.