Auditory Warning Signal Research Articles

Psychological evaluation of auditory warning signal for train horn

Air horns are commonly used to signal danger or caution during railroad operations. The main objectives of those auditory signals (i.e., whistle cords) are to "warn of danger," "call attention," and "notification of approach" in the situation before an accident occurs or "report an emergency" to those around after an accident occurs. Further, in case of using auditory signals in nonemergency situations (e.g., during the departure or arrival of a train), loud or unpleasant sounds should be avoided to the greatest possible extent, considering noise pollution in the environment. This study examined the nature of auditory signals for railroad vehicles through psycho-acoustic experiments using 11 types of auditory stimuli, including a "short, rapid horn sound" and a "long, slow horn sound." In the rating experiments, the auditory stimuli were presented via headphones to individual participants. The results indicated that upon initiating the auditory signal with an electronic horn sound (less unpleasant and dangerous impression) before blowing the air horn to alert perceived danger, the impressions of "warning," "caution," and "danger" tended to decrease, whereas those of "pleasantness," "quietness," and "relaxedness" tended to increase. The combined usage of electric and air horns can effectively provide auditory signals for maintaining the safety of railroad operations while addressing the associated noise pollution.

On the selection of endogenous and exogenous signals.

Alerting is one of the three components of attention which involves the eliciting and maintenance of arousal. A seminal study by Posner et al. (Posner MI, Klein R, Summers J, Buggie S. 1973 Mem. Cognit. 1, 2–12 (doi:10.3758/BF03198062)) focused on how changing the interval between an alerting signal and a target would impact the speed and accuracy of responding. Participants indicated whether targets were presented on the left or right side of the fixation point. Auditory warning signals were played at various intervals prior to the target to alert participants and prepare them to make a response. Reaction times revealed a robust, U-shaped, preparation function. Importantly, a clear speed-accuracy trade-off (SAT) was observed. In the current experiment, we replicated the methodological components of this seminal study while implementing a novel auditory warning signal (Lawrence MA, Klein RM. 2013 J. Exp. Psychol. General 142, 560 (doi:10.1037/a0029023)) that was either purely endogenous (change in quality without a change in intensity; analogous to isoluminant colour change in vision) or a combination of endogenous and exogenous (change in both quality and intensity). We expected to replicate the U-shaped preparation function and SAT observed by Posner and colleagues. Based on Lawrence and Klein's findings we also expected the SAT to be more robust with the intense signal in comparison to the isointense signal.

Active noise control system in a fire helmet – a proposal of the solution

An auditory warning signal emitted by a moving emergency vehicle is intended to inform other road users about an approaching emergency vehicle and the need to give way to it. Emission of a warning signal with the highest possible sound pressure level is the main method of ensuring the audibility and recognition of the warning signal by traffic participants at unknown and varying traffic noise, acoustic insulation of vehicles and noise in the interior of the vehicle. The A-weighted sound pressure level of the auditory warning signal reaching the interior of an emergency vehicle may exceed 90 dB, which can be harmful to the hearing, can adversely affect psychophysical fitness of the crew members and can significantly hinder verbal communication in the vehicle. This article presents the concept and results of numerical simulations of an active noise control system in fire helmets, which will reduce the exposure of the crew to the warning signal and improve the quality of verbal communication in an emergency vehicle. This system is integrated with the signal generation system. The Notch algorithm and the NLMS adaptation algorithm were used in the system's control algorithm. The construction of the controller was based on a microcontroller from the STM32F4 family.

Postural control and contingent negative variation during transient floor translation while standing with the ankle fixed

BackgroundAdaptation changes in postural muscle activity and anticipatory attention were investigated with the ankle joint fixed to change postural control strategies during transient floor translation.MethodsFor 15 healthy young adults, 40 transient floor translations (S2) in the anterior direction were applied 2 s after an auditory warning signal (S1), under conditions with or without fixation of the ankle. Activity of the frontal postural muscles (tibialis anterior (TA), rectus femoris (RF), rectus abdominis) and contingent negative variation (CNV, brain potential) were analyzed for 20 trials each of the early and latter halves under each fixation condition.ResultsWith fixation, peak amplitude of muscle activity after S2 was significantly decreased in TA and increased in RF. These muscles showed marked adaptive decreases. The early component of CNV reduced with adaptation, particularly under fixation condition. Only in RF, background activity increased just before S2, with adaptation under fixation. A significant correlation was found between timings of CNV peak and RF activation just before S2 only after adaptation under fixation.ConclusionThese results suggest that the main activation muscle changes from TA to RF with fixation. Under such condition, attention would be focused on the knee with adaptation, and the need for heightening attention in the early stage may have declined. Correspondingly, the timing to heighten stiffness of the RF became later, and attention would have been paid to RF activation just before S2.

Warning Signals With Rising Profiles Increase Arousal

Auditory warnings are often used to direct a user’s attention from a primary task to critical peripheral events. In the context of traffic, in-vehicle collision avoidance systems could, for example, employ spatially relevant sounds to alert the driver to the possible presence of a crossing pedestrian. This raises the question: What is an effective auditory alert in a steering environment? Ideally, such warning signals should not only arouse the driver but also result in deeper processing of the event that the driver is being alerted to. Warning signals can be designed to convey the time to contact with an approaching object (Gray, 2011). That is, sounds can rise in intensity in accordance with the physical velocity of an approaching threat. The current experiment was a manual steering task in which participants were occasionally required to recognized peripheral visual targets. These visual targets were sometimes preceded by a spatially congruent auditory warning signal. This was either a sound with constant intensity, linearly rising intensity, or non-linearly rising intensity that conveyed time-to-contact. To study the influence of warning cues on the arousal state, different features of electroencephalography (EEG) were measured. Alpha frequency, which ranges from 7.5 to 12.5 Hz, is believed to represent different cognitive processes, in particular arousal (Klimesch, 1999). That is, greater desynchronization in the alpha frequency reflects higher levels of attention as well as alertness. Our results showed a significant decrease in alpha power for sounds with rising intensity profiles, indicating increased alertness and expectancy for an event to occur. To analyze whether the increased arousal for rising sounds resulted in deeper processing of the visual target, we analyzed the event related potential P3. It is a positive component that occurs approximately 300 ms after an event and is known to be associated with recognition performance of a stimulus (Parasuraman & Beatty, 1980). In other words, smaller P3 amplitudes indicate worse identification than larger amplitudes. Our results show that sounds with time-to-contact properties induced larger P3 responses to the targets that they cued compared to targets cued by constant or linearly rising sounds. This suggests that rising sounds with time-to-contact intensity profiles evoke deeper processing of the visual target and therefore result in better identification than events cued by sounds with linearly rising or constant intensity.

Aging effect on detectability, criticality and urgency under various auditory conditions

The objective of the study was to investigate the relationship between different components of an auditory warning signal and determine elderly driver’s detection, criticality and urgency perceptions under varied auditory warning conditions. A laboratory study was conducted involving 28 younger and 28 older adults. A pure-tone audiogram was administered for quantification of the participant’s hearing level (in dBHL) prior to the test protocols. Audiogram was conducted in a sound isolated booth. The main experiment investigated the effects of frequency, temporal characteristics, aging, and noise on detectability, caution and perceived urgent levels. Method of adjustment was used by the subjects to adjust the sound pressure level (dBA) till they detected, or felt caution or urgent. Overall, the results indicated that elderly had higher sound pressure threshold than younger individuals in all the sessions (i.e., detectability, caution and urgency). Elderly were found to have higher hearing thresholds especially at 4000Hz. Higher frequencies were also seen to require less sound pressure thresholds to convey feeling of caution and urgency. Sound pressure threshold to convey the feeling of urgency was the lowest for 2000Hz and 4000Hz with 1000Hz requiring the highest threshold intensity. In conclusion, current study findings suggest frequency, temporal and spectral components to influence a person’s criterion. We would recommend the use of complex tones (multi-components) having high frequency around 2000Hz for detection, warning and urgent conditions.

Postural control during transient floor translation while standing with the leg and trunk fixed

Postural movement was restricted above the ankle, and contingent negative variation (CNV) and postural muscle activity were investigated during a transient floor translation (S2) 2s after an auditory warning signal (S1). For 13 healthy young adults, the joints of the knee, hip, and trunk were fixed using a cast brace. Under no-fixation and fixation, a set of 10 translations was repeated at least 4 times, and center of pressure in the anteroposterior direction (CoPap), posterior postural muscle activity of the body (elector spinae (ES), biceps femoris (BF), gastrocnemius (GcM) and soleus (Sol)), and late CNV at Cz were analyzed in the initial two sets (initial set) and last two sets (late set). In the no-fixation, CoPap forward displacement after S2 gradually decreased. In the first trial of the fixation, it had significantly increased, and then rapidly decreased across subsequent trials. CNV peak amplitude was largest in the late set of the fixation. The activity of postural muscles increased just before S2 and in the late set the start time showed high correlations with CNV peak time in all muscles (ES (r=0.88), BF (0.92), GcM (0.80), and Sol (0.89)) under the no-fixation, but exclusively in the GcM (0.84) and Sol (0.88) under the fixation. When postural control was restricted mainly to the ankle, attention would have been focused mainly on processing sensory information from the triceps surae just before the floor translation.

Adaptation changes in dynamic postural control and contingent negative variation during repeated transient forward translation in the elderly

BackgroundAdaptation changes in postural control and contingent negative variation (CNV) for the elderly were investigated during repeated forward floor translation.MethodsFifteen healthy elderly persons, living in the suburban area of Kanazawa City, Japan, underwent backward postural disturbance by a forward-floor translation (S2) 2 s after an auditory warning signal (S1). A set with 20 trials was repeated until a negative peak of late CNV was recognized in the 600-ms period before S2, and the last set was defined as the final set. Electroencephalograms, center of foot pressure in the anteroposterior direction (CoPap), and electromyograms of postural muscles were analyzed.ResultsCoPap displacement generated by the floor translation was significantly decreased until the twelfth trial in the first set, and mean CoPap displacement was smaller in the second and final sets than in the first set. The mean displacement was significantly smaller in the final set than the previous set. A late CNV with a negative peak was not recognized in the first and second sets. However, most subjects (13/15) showed a negative peak by the fourth set, when the late CNV started to increase negatively from about 1,000 ms after S1 and peaked at about 300 ms before S2. At about 160 ms before the CNV peak, the CoPap forward shift started. The increase in timing of the gastrocnemius activity related to the CoPap shift was significantly correlated with the CNV peak timing (r = 0.64). After S2, peak amplitudes of the anterior postural muscles were significantly decreased in the final set compared to the first set.ConclusionsIt was demonstrated that even for the elderly, with so many repetitions of postural disturbance, a late CNV with a negative peak was recognized, leading to accurate postural preparation. This suggests the improvement of frontal lobe function (e.g., anticipatory attention and motor preparation) in the elderly.

Influence of anxiety, depression and looming cognitive style on auditory looming perception

Previous studies show that individuals with an anticipatory auditory looming bias over-estimate the closeness of a sound source that approaches them. Our present study bridges cognitive clinical and perception research, and provides evidence that anxiety symptoms and a particular putative cognitive style that creates vulnerability for anxiety (looming cognitive style, or LCS) are related to how people perceive this ecologically fundamental auditory warning signal. The effects of anxiety symptoms on the anticipatory auditory looming effect synergistically depend on the dimension of perceived personal danger assessed by the LCS (physical or social threat). Depression symptoms, in contrast to anxiety symptoms, predict a diminution of the auditory looming bias. Findings broaden our understanding of the links between cognitive-affective states and auditory perception processes and lend further support to past studies providing evidence that the looming cognitive style is related to bias in threat processing.

Adaptation changes in dynamic postural control and contingent negative variation during backward disturbance by transient floor translation in the elderly

BackgroundWe investigated adaptation changes in dynamic postural control and contingent negative variation (CNV) in 13 young and 12 elderly adults. Subjects repeatedly underwent backward postural disturbance by a forward floor translation (S2) 2 s after an auditory warning signal (S1). Initial and second sets were conducted, each set with 20 trials. Posterior peak position of the center of pressure in the anteroposterior direction (CoPy) after S2 was identified. Electroencephalograms from Cz were averaged for each set, and the CNV negative peak was identified.ResultsCompared with the first trial, the posterior peak position of CoPy changed significantly forward from the 12th trial in the young and from the 19th trial in the elderly during the initial set. The mean of the posterior peak position was more forward in second set than in the initial set for both groups and was significantly backward in the elderly compared to the young for both sets. These findings indicate that subjects in both groups adapted better to the postural disturbance in the second set than in the initial set, and the adaptation was later in the elderly. Late CNV in the young started to increase negatively from the middle of the S1-S2 period and peaked just before S2. Peak CNV amplitude was larger in the second set than in the initial set. In contrast, late CNV in the elderly exhibited no negative increase as in the young and peaked in the middle of the S1-S2 period, which was followed by gradual decreasing toward S2. No adaptive changes were found in late CNV for the elderly.ConclusionsIt is conceivable that reduced activation of the frontal lobe may be one of the factors contributing to the decrease in postural adaptability in the elderly. The elderly may use various brain regions for the adaptation of dynamic postural control compared with the young.

The effect of a cross-trial shift of auditory warning signals on the sequential foreperiod effect

When a warning signal (WS) precedes an imperative signal (IS) by a certain amount of time (the foreperiod, FP), responses are speeded. Moreover, this effect is modulated by the FP length in the previous trial. This sequential FP effect has lately been attributed to a trace-conditioning mechanism according to which individuals learn (and re-learn) temporal relationships between the WS and the IS. Recent evidence suggests that sensory WS attributes are critical to trigger time-related response activation. Specifically, when WS modality is shifted in subsequent trials (e.g., from auditory to visual modality), the sequential FP effect becomes attenuated. This study examined whether the sequential FP effect is reduced only by between-modality shifts or whether this attenuation generalizes to cross-trial shifts of WS attributes within modalities. We compared dimensional (low vs. high tone frequency) and qualitative shifts (pure tone vs. noise) of equal-intense auditory WS events. The results of four experiments revealed that shifts of tone frequency did not, whereas shifts of qualitative tone characteristics did attenuate the sequential FP effect. These results support the view that the WS acts as a trigger cue that unintentionally activates responses at previously reinforced critical moments.

Using Peripersonal Warning Signals to Orient a Driver’s Gaze

We report a series of three experiments designed to assess the relative speed with which people can initiate speeded head-orienting responses following the presentation of spatial warning signals. Recent cognitive neuroscience findings have shown that the human brain tends to treat stimuli occurring in peripersonal space as being somehow more behaviorally relevant and attention demanding than stimuli occurring in extrapersonal space. These brain mechanisms may be exploited in the design of warning signals. Experiment 1 assessed the effectiveness of various different unisensory warning signals in eliciting a head-turning response to look at the potential source of danger requiring participants' immediate attention; Experiment 2 assessed the latency of a driver's responses to events occurring in the cued direction; Experiment 3 assessed the relative effectiveness of various warning signals in reorienting a person's gaze back to a central driving task while he or she was distracted by a secondary task. The results show that participants initiated head-turning movements and made speeded discrimination or braking responses significantly more rapidly following the presentation of a close rear auditory warning signal than following the presentation of either a far frontal auditory warning signal, a vibrotactile warning signal presented to their waist, or a peripheral visual warning signal. These results support the claim that the introduction of peripersonal warning signals results in a significant performance advantage relative to traditionally designed warnings. Warning systems that have been designed around the constraints of the human brain offer great potential in the future design ofmultisensory interfaces.

Centrifugal modulation of human LEP components to a task-relevant noxious stimulation triggering voluntary movement

In the present study, we investigate the top-down centrifugal modulation of neural responses to a task-relevant noxious stimulation triggering voluntary movement by recording magnetoencephalography (MEG) and electroencephalography (EEG) simultaneously. An auditory warning signal was followed 2–3 s later by a noxious YAG laser stimulation as an imperative signal delivered to the left hand dorsum. Ten normal subjects performed three different conditions, Control, Movement, and Count. In Control, the subjects were asked to relax and rest quietly with no task. In Movement, the subjects extended the left index finger after imperative stimuli. In Count, the subjects counted the number of imperative stimuli silently. The amplitude of the N2 component recorded by EEG, which peaked about 220 ms after noxious stimulation, was significantly attenuated in Movement, but not in Count, compared to Control. The root-mean-square (RMS) from both hemispheres, and areal mean signal (AMS) amplitudes and the equivalent current dipole (ECD) strengths from SI/PPC and bilateral SII recorded at around 170 ms by MEG were not significantly different among the three conditions. In contrast, ECD strengths and AMS amplitudes from the anterior cingulate cortex (ACC), which showed a similar peak to the N2 component, were smaller in Movement than Control and Count. We therefore suspect that neural activities related to generator mechanisms of N2, especially including ACC, are inhibited by movement-related neural activities during the preparatory period. The present findings indicate a characteristic of pain–motor integration in a movement preparatory period.

The modality shift effect and the effectiveness of warning signals in different modalities

Which is better, a visual or an auditory warning signal? Initial findings suggested that an auditory signal was more effective, speeding reaction to a target more than a visual warning signal, particularly at brief foreperiods [Bertelson, P., & Tisseyre, F. (1969). The time-course of preparation: confirmatory results with visual and auditory warning signals. Acta Psychologica, 30. In W.G. Koster (Ed.), Attention and Performance II (pp. 145–154); Davis, R., & Green, F. A. (1969). Intersensory differences in the effect of warning signals on reaction time. Acta Psychologica, 30. In W.G. Koster (Ed.), Attention and Performance II (pp. 155–167)]. This led to the hypothesis that an auditory signal is more alerting than a visual warning signal [Sanders, A. F. (1975). The foreperiod effect revisited. Quarterly Journal of Experimental Psychology, 27, 591–598; Posner, M. I., Nissen, M. J., & Klein, R. M. (1976). Visual dominance: an information-processing account of its origins and significance. Psychological Review, 83, 157–171]. Recently [Turatto, M., Benso, F., Galfano, G., & Umilta, C. (2002). Nonspatial attentional shifts between audition and vision. Journal of Experimental Psychology: Human Perception and Performance, 28, 628–639] found no evidence for an auditory warning signal advantage and showed that at brief foreperiods a signal in the same modality as the target facilitated responding more than a signal in a different modality. They accounted for this result in terms of the modality shift effect, with the signal exogenously recruiting attention to its modality, and thereby facilitating responding to targets arriving in the modality to which attention had been recruited. The present study conducted six experiments to understand the cause of these conflicting findings. The results suggest that an auditory warning signal is not more effective than a visual warning signal. Previous reports of an auditory superiority appear to have been caused by using different locations for the visual warning signal and visual target, resulting in the target arriving at an unattended location when the foreperiod was brief. Turatto et al.’s results were replicated with a modality shift effect at brief foreperiods. However, it is also suggested that previous measures of the modality shift effect may still have been confounded by a location cuing effect.

The course of neuropsychological functions in children with attention deficit hyperactivity disorder from late childhood to early adolescence

The aim of this follow-up study was to investigate the course of performance in attentional tasks in children with ADHD and normal controls in late childhood and preadolescence over short periods of time. The development of two dimensions of attention was compared: alertness/arousal and inhibitory control. Children with ADHD (N=28) and normal controls (N=25) were examined at three times: at baseline (age mean=10.8 years, SD=1.5), after one year (age mean=12.0 years, SD=1.6), and after 2.6 years (age mean=13.3 years, SD=1.6). They performed two tasks of a computerized battery for attentional performance: Alertness--a test of simple reaction time to visual stimuli contrasting a condition with and without auditory warning signal, and Incompatibility--a test of spatial interference/inhibitory control. Clinical diagnosis according to DSM-III-R criteria was established at time 1 and time 3 by structured diagnostic interviews. In the Alertness task significant group differences regarding increased reaction time variability in ADHD, but not for reaction time itself, were found at time 1 and more pronounced at time 2. At time 3 group differences had disappeared. In the Incompatibility task group differences in number of errors were not observed at time 1, whereas children with ADHD made significantly more errors at time 2 and less pronounced at time 3. The degree of clinical symptom remission after 2.6 years was not related to changes in neuropsychological performance. When measuring attentional functions, the selection of an appropriate time window seems to be essential for the detection of group differences between ADHD children and controls, because group differences are most pronounced before adolescence. The different developmental course of selective components of attention should be taken into account.

Conjunction in simulated railway signals: a cautionary note

Train drivers routinely perform visual search tasks to locate combinations of coloured signals controlling their progress, and are required to make discrete decisions on the basis of what they see. Two studies are reported which examine the performance of students under conditions that simulate critical aspects of United Kingdom train drivers' signal-response task. The most crucial cautionary signal, the single yellow signal used to alert a transition to potentially hazardous situations, was responded to more slowly than other signal types. A longer processing time was found whether (Study 2) or not (Study 1) the signal appearance was accompanied by the auditory warning signal train drivers encounter under actual driving conditions. The results are consistent with predictions from Treisman and Gelade's (1980) Feature Integration Theory, and the implications for signal sighting practice are discussed. Copyright © 2005 John Wiley & Sons, Ltd.

Stimulus array onset as a preparatory signal in attentional selection.

Warning signals enhance psychomotor performance by optimising preparation for the arrival of an event. Recent evidence, however, suggests that a warning signal can also disrupt attentional preparation by interfering with the process of preparing. It was hypothesised that a warning signal consisting of a change to the task-relevant items (array onset) may be more effective than a traditional warning signal consisting of the arrival, or removal, of a bar-cue which is independent of the task array. In three experiments array onset was a more effective warning signal than the bar-cue because reaction times were significantly faster without an increase in errors. In addition, an auditory warning signal resulted in faster reaction time than the bar-cue but in performance equivalent to that with an onset warning signal. Thus, the superiority of an auditory warning signal, reported by Davis and Green, was not found, when the interference of the visual warning signal with preparation was reduced. These results suggest that a visual warning signal consisting of a change to the stimulus array is more effective than an event independent from the stimulus array.

Temporal Preparation Strategy may Inflate RT Deficit in Patients With Parkinson’s Disease

Twelve patients with Parkinson’s disease (PD) and 12 age-matched controls completed a visual reaction time (RT) task to assess the effect of temporal parameters on response preparation. Simple and choice RT conditions were presented in separate blocks. In both conditions, preparatory intervals of various durations (1, 3 and 5 s) were introduced between an auditory warning signal and the visual target. Within a block of trials, intervals varied randomly. The results indicated that PD patients responded slower than controls in both task conditions. Also, there was evidence for preparation in both groups, as RT decreased with increasing intervals. A three-way interaction indicated that PD patients’ RT was longer than that of controls at the shortest interval in simple RT. This suggests that PD patients show a different pattern of temporal response preparation and that this may contribute to their deficit on RT tasks.

Development of voluntary control of saccadic eye movements

To investigate the development of the voluntary control of saccadic eye movement, we examined eye movements in 99 normal children (4-13 years of age). Subjects were asked to fixate a central light for 3-5 s. A target was then presented, either to its right or left. In visually guided saccades, the mean latencies of the child group were longer than those of the adults, and decreased with age until the age of 12 where they reached adult levels. On the other hand, their peak saccadic velocities were not different from those of the adults. In the antisaccade task, they showed a higher rate of directional errors, indicating an inability to suppress reflexive saccades to the target. Mean latencies of correct antisaccades were significantly longer in the children than in the adults. Error rates and antisaccadic latency tended to decrease with age. We also examined the effects of an auditory warning signal during the fixation period and compared the results with those without. The warning stimulus was less effective in children than in adults in both tasks. Similar peak saccadic velocities between children and adults suggest the earlier development of the saccadic burst generator in the brainstem. In contrast, the delayed development of latency of saccades and antisaccades and the error rates of antisaccades suggest delayed maturation of the cerebral cortex, particularly the frontal association areas that are involved in both eye movement control and attentional processes.

Estimating the components of the gap effect.

The gap effect refers to the finding that saccadic latencies are typically reduced when a fixation point is removed prior to the appearance of a peripheral target. This reduction in saccadic reaction time (SacRT) is thought to be due to a general warning effect and an oculomotor specific fixation offset that occur when the fixation point is removed. In order to estimate the contribution of each of these effects to the overall gap effect, this paper introduces a new manipulation, the partial-gap trial, where the fixation point undergoes a change in size prior to the presentation of the target. The partial-gap trial is presumed to provide the visual warning effect of the fixation offset (i.e. similar to that in a gap trial) but does not provide the fixation offset effect (FOE). When the fixation point was abruptly reduced in size before the presentation of the target, the estimated decrease in SacRT due to the visual warning effect was 5-7% and did not differ in the presence or absence of an auditory warning signal. It was found that auditory warning effect and the FOE interacted in reducing SacRTs. Additionally, when the fixation point was abruptly increased in size before the presentation of the target, SacRTs were slower than when the fixation point did not change in size and remained present for the entire trial (i.e. an overlap trial). We conclude that this new partial-gap paradigm is a useful method for researchers wishing to separately examine FOE and visual warning effects.