- Email: [email protected]
Conditioned facilitation of the unconditioned reflex after classical eyeblink conditioning
Available online at www.sciencedirect.com
International Journal of Psychophysiology 67 (2008) 17 – 22 www.elsevier.com/locate/ijpsycho
Conditioned facilitation of the unconditioned reflex after classical eyeblink conditioning Ole Åsli ⁎, Magne Arve Flaten Department of Psychology, University of Tromsø, N-9037 Tromsø, Norway Received 13 April 2007; received in revised form 18 September 2007; accepted 20 September 2007 Available online 4 October 2007
Abstract The present study (N = 40) investigated the time-course of conditioned facilitation of the unconditioned eyeblink reflex (UR). In a single-cue delay classical conditioning procedure, a tone conditioned stimulus (CS) signaled an airpuff unconditioned stimulus (airpuff US) to the eye. A paired group received 40 trials of CS/US presentations with an interstimulus interval (ISI) of 200 ms. An unpaired group received an equal number of explicitly unpaired presentations of the CS and US. Thereafter, eyeblink reflex facilitation was assessed by presenting 94 dB white noise (noise US) 10, 30, 50, 100, 150, and 1000 ms after CS onset. In the paired group, URs were significantly increased as early as 100 ms after CS onset compared to the unpaired group. This reflex facilitation was correlated with CR magnitude, indicating that conditioned facilitation of eyeblink URs indexes an early, automatic, preattentive stage in CR formation. © 2007 Elsevier B.V. All rights reserved. Keywords: Classical conditioning; Eyeblink conditioning; Conditioned facilitation; Reflex facilitation
1. Introduction Several studies have found that unconditioned reflexes (URs), elicited in the presence of a conditioned stimulus (CS), are facilitated compared to URs elicited in the absence of the CS (Flaten 1993; Flaten and Hugdahl 1991; Weisz and McInerney 1990; Wikgren et al. 2002). The present study investigated the time-course of conditioned facilitation of URs. It has been hypothesized that conditioned facilitation of the UR indexes an early stage of eyeblink CR formation (Flaten 1993; Weisz and McInerney 1990). Flaten (1993) used an airpuff to the eye as the US, and found eyeblink UR facilitation to a noise US after only four CS/US pairings, prior to the occurrence of eyeblink conditioned responses (CRs). Similar findings were made by Weisz and McInerney (1990) and Wikgren et al. (2002). It was argued that conditioned UR facilitation was the result of a developing associative process that would eventually produce the eyeblink CR. The process underlying UR facilitation could be CS-induced facilitation of the motoneuron pool in the facial ⁎ Corresponding author. Tel.: +47 776 46778; fax: +47 776 45610. E-mail address: [email protected] (O. Åsli). 0167-8760/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpsycho.2007.09.003
nerve that is responsible for the eyeblink (e.g., Guyton and Hall, 2006). After CS presentation, it is hypothesized that the motoneuron pool is gradually more facilitated until a CR is elicited or not. Consequently, when a reflex-eliciting stimulus is presented to an already facilitated motoneuron pool, the magnitude of the reflex increases. The degree of increase of the reflex would depend on the number of facilitated neurons and the amount of facilitation or depolarization. There is disagreement on the role of voluntary responding and awareness of the CS/US contingency in eyeblink classical conditioning. (Clark and Squire 2004; Manns et al. 2001; Smith et al. 2005) argued that single-cue delay classical conditioning is independent of awareness of the CS/US contingency. However, their measure of awareness was criticized by Lovibond and Shanks (2002a) for underestimating awareness of the CS / US contingency. Lovibond and Shanks (2002b) argued that the claim that automatic, non-conscious processes played a role in human conditioning had not been proven. Rather, a conscious, propositional system could explain the development of the CR. One central point in this debate is the claim by Lovibond and Shanks (2002b) that almost every study of the role of consciousness in classical conditioning used inadequate measures of
18
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
awareness, and that most measures tend to underestimate awareness, or that the experimental procedures used may lead to underestimated awareness. The present study bypassed the problem of awareness and voluntary responding by investigating CR and UR facilitation at points in time after CS onset that were too short for conscious processes to play a role. The participants in the paired group first received 40 trials of eyeblink conditioning with a short interstimulus interval (ISI) of 200 ms. The short ISI was used to reduce the role of voluntary responding in the CR data. In the unpaired group, an equal number of CSs and USs were presented in an explicitly unpaired manner. Thereafter, UR facilitation was tested by presenting eyeblink reflex-eliciting noise immediately after CS presentation. The stimulus onset asynchronies (SOAs) between CS onset and noise onset were chosen to investigate the latency and development of the process of UR facilitation. A positive correlation between CR amplitude and UR facilitation was expected if the latter could be related to a process underlying CR production. Evaluations of CS and US valence and arousal were obtained to investigate whether emotional processes could be related to UR facilitation. 2. Materials and methods 2.1. Subjects Forty subjects (23 females and 17 males, age range 19–37, mean age 25.9 years) participated in the study. The participants were randomly assigned to one of two groups (paired group or unpaired group), each containing 20 participants. All participants were in good health and did not report any previous serious disease or injury. All participants had auditory thresholds of 30 dB or less at 1000 Hz in both ears. The project was approved by the Regional Committee of Medical Research Ethics in Health Region IV in Norway (Project 29/2000). The participants were instructed to refrain from drinking caffeinated beverages and use nicotine-containing substances for 3 hours prior to the start of the study. Written informed consent was obtained from all participants, who where paid 2 lottery tickets (equivalent to 50 NOK) for their participation. 2.2. Apparatus and stimuli The experiment took place in an electrically and acoustically shielded chamber (C. A. Tegnèr), where the temperature was kept at about 20 °C. Hearing was tested with a Grason-Stadler 17 audiometer. A Bruel and Kjær 2235 Sound Level Precision Meter measured intensity of auditory stimuli. Programs written by the last author in ASYST 3.1 and run on an ALR 486 PC, controlled presentation of experimental stimuli and data acquisition. There were two different unconditioned stimuli (US) in the experiment. The airpuff US was paired with the conditioned stimuli (CS). The noise US was deployed to elicit eyeblink UR and was not paired with a CS. The noise US had an intensity of 94 dB, instantaneous rise time and a duration of 50 ms. The CSs were 1000-Hz tones with intensity of 60 dB, and a rise time of 20 ms. The duration of the prepulses were the same as the
stimulus onset asynchrony (SOA) between the tone prepulse and the eyeblink-eliciting noise. The stimuli were delivered through Sennheiser HD 250 headphones. The airpuff US was directed to the corner of the right eye. The airpuffs had an intensity of 20 kPa, as indicated at the pressure regulator on the air-pressure bottle, and a duration of 50 ms. The airpuffs were delivered by apparatus described in Flaten et al. (1989). Eyeblink electromyographic (EMG) responses were recorded from the right orbiculari oculi with two Ag/AgCl Sensor Medics miniature electrodes (2 mm diameter) filled with Ultra Phonic conductivity gel. Inter-electrode distance was 1.5– 2.5 cm. The ground electrode was placed centrally on the forehead. The EMG signal was amplified by a factor of 60,000 and filtered (passing 90–250 Hz) by a Coulbourn S75-01 bioamplifier. The signal was rectified and integrated with a Coulbourn S76-01 contour-following integrator with a 10-ms time constant, and the output sent to the PC via a Keithley interface. Sampling on each trial began 200 ms prior to onset of the first stimulus and continued for 200 ms after onset of the US. The sampling rate was 10 Hz prior to onset of the first stimulus and 1000 Hz after stimulus onset. Emotional valence and arousal elicited by the CS and airpuff US was recorded with the Self Assessment Manikin (SAM) (Bradley and Lang 1994). General stress and arousal was recorded with the O'Neill and Parrott (1992) stress and arousal scale. 2.3. Procedure After arrival at the laboratory the subjects sat down in a desk chair and read and filled in the Informed Consent Form. Level of hearing was tested before the participants were lead into the experimental chamber and seated in a reclining chair. The subjects were informed of the general purpose of the study, about the stimuli and procedure, and that they could withdraw from the study without giving any reason at any time. The skin below the participants' right eye was cleaned with a swab containing alcohol and pumice, and the electrodes for measurement of the eyeblink electromyography (EMG) were attached. The headgear containing the airpuff-presentation equipment and the headphones were attached, and then the subjects filled out the O'Neill and Parrott (1992) stress and arousal scale. The door to the experimental chamber was closed, and after a brief pause the experimental procedure was in effect. The experiment consisted of two phases (see Fig. 1). In the conditioning phase, the paired group received 40 trials of singlecue classical conditioning with the 1000 Hz tone as the CS and the airpuff to the eye as the US. The interstimulus interval (ISI) from CS onset to US onset was 200 ms. The unpaired group received 40 explicitly unpaired presentations of the CS and US. In the noise UR phase, UR facilitation to the noise US was investigated by presenting the CS at SOAs of 10, 30, 50, 100, 150, and 1000 milliseconds relative to onset of the eyeblink eliciting noise. Each SOA was presented four times. The noise US was also presented four times alone, so a total of 28 trials were presented. The SOAs were presented in semi-random
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
19
onset. Thus, alpha responses could not be scored mistakenly as CRs. To ascertain this, a separate analysis of alpha responses was performed that confirmed that there was no difference between the paired group and the unpaired group in alpha responses during conditioning. 2.5. Design and statistics The design for the conditioning phase was a 2 Conditioning (paired, unpaired) × 8 Trial blocks mixed design where the first factor was treated as a between-subjects factor and the last factor was treated as a within-subjects factor. The design for the noise UR phase of the experiment was a 2 Conditioning (paired, unpaired) × 7 SOA (noise alone, 10, 30, 50, 100, 150, and 1000 ms) mixed design with the first factor treated as a between-subjects factor and the last factor treated as a within-subjects factor. The data were analysed with repeated measures analysis of variance (Statistica 7.0). Theoretically interesting contrasts were examined by contrast analyses. 3. Results 3.1. Subjective data There were no significant main effects or interactions in the ratings of CS and US valence and arousal. 3.2. Conditioned responses Fig. 1. Schematic presentation of the procedure.
order in two blocks of 12 trials each. The airpuff US was presented 400 ms after CS onset at 20 trials in the paired group in this phase to avoid extinction. The airpuff US was not presented on trials where the SOA was 1000 ms. At the end of the experiment the subjects filled in the O'Neill and Parrott posttest, and two SAM forms for recording of affective reactions to the CS and airpuff US.
There was a significant main effect of Conditioning (F(1,38) = 13.23, p = .0008), due to significantly larger conditioned responses in the paired group compared to the unpaired group. There was a significant interaction of Conditioning by Trial block (F(7,266) = 3.02, p = .004), and follow-up tests showed significantly stronger CRs in the paired group compared to the unpaired group in Trial blocks 3, 4, 5, 6, 7 and 8 (all ps b .006) (Fig. 3.). The largest difference in CR magnitude between the groups was 5.9 μV in Trial block 6.
2.4. Response scoring and data reduction Noise unconditioned eyeblink reflexes were scored as the difference between the mean of a 200 ms baseline recorded prior to the first stimulus presented in that trial, and the maximum EMG level in the interval from 20 to 120 ms after onset of the noise. Noise URs were also expressed as proportion of difference from control: Average responding to each SOA was calculated across the entire session. Average responding to noise US alone (control) was subtracted from the average responses to each SOA and this difference was divided by the average responding to the noise US alone (Blumenthal et al., 2004). Conditioned eyeblink reflexes (CRs) were scored as the difference between the mean of a 200 ms pre-stimulus baseline and the maximum EMG level in the interval from 130 to 200 ms after CS onset. This window was chosen since alpha responses, i.e., unconditioned responses occasionally occurring after CS onset, are usually terminated between 100 and 120 ms after CS
Fig. 2. Left panel: Conditioned eyeblink responses in the paired and unpaired groups in the conditioning phase. Error bars represent 1 standard error of the mean. Right panel: Noise UR magnitudes across SOAs. S2 denotes noise alone trials.
20
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
Fig. 3. Grand average of responses to the noise US alone and to the noise US presented 100 ms after the CS, in the paired and unpaired groups.
3.3. Noise unconditioned response magnitude There was no difference between the paired and the unpaired group in noise UR magnitude to noise alone (F b 1, ns). The main effect of SOA was significant (F(6, 228) = 31.08, p = .00001). There was also a significant interaction of Conditioning by SOA (F(6, 228) = 11.80, p = .000001) and follow-up tests showed significantly more facilitation of noise URs elicited 100 and 150 ms after CS onset in the paired group compared to the unpaired group ( p = .004 and p = .0002, respectively) (Fig. 2). The greatest difference in noise UR magnitude between the paired and the unpaired group was at 150 ms, where there was a difference of about 14.9 μV. Fig. 3 shows the mean EMG for each millisecond for all acceptable trials for the noise-elicited Table 1 Correlations between CRs in the first 20 trials (First) and the last 20 trials (Last) in the conditioning phase, and noise-elicited eyeblink URs across the SOAs in the noise UR phase
S2 10 30 50 100 150 1000 S2 denotes noise alone trials. a p b .05.
First
Last
0.11 0.17 0.16 0.17 0.41 a 0.46 a 0.29
0.24 0.27 0.24 0.30 0.56 a 0.55 a 0.37 a
UR to the noise alone, and for noise-elicited URs to the noise presented 100 ms after CS onset, i.e. at the shortest SOA where significant reflex facilitation was observed. Noise-elicited URs expressed as proportion of difference from control revealed similar results as UR magnitude analyses. 3.4. Correlations between CR magnitude and UR magnitude There were significant positive correlations between CRs in the second half of the trials in the Conditioning phase, and noise URs at the 100, 150, and 1000 ms SOAs in the Noise UR phase (Table 1). 4. Discussion There was a significant interaction of Conditioning by SOA in the eyeblink UR data, due to significantly increased URs at the 100 and 150 ms SOAs in the paired group compared to the unpaired group. This short latency of facilitated UR indicates that the mediating process is automatic and pre-attentive. As attentional modulation of URs has been seen at short SOAs (Elden and Flaten 2002, 2003; Filion and Poje 2003), such an interpretation is worth mentioning. Elden and Flaten (2002) compared task and no-task conditions presented in separate blocks, and found significantly greater inhibition of the UR in the task than no-task condition at SOAs between 30 and 420 ms (Experiment 2). This led the researchers to conclude that preparatory attention, i.e., attention to a stimulus about to occur,
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
inhibited subsequently elicited UR. Preparatory attention to the CS could have been induced in the participants in the present study. However, preparatory attention cannot explain the present results, since attention to the CS in the paired group would have inhibited the UR, i.e., the opposite of the facilitated the UR seen in the present experiment. The UR facilitation seen in the paired group was most likely not CRs mistakenly scored as URs to the noise. The largest difference in CRs between the paired and unpaired groups in the conditioning phase was 5.9 μV in trial block six, a relatively small difference compared to the difference of 14.9 μV in UR facilitation in the noise UR phase. Thus, it is unlikely that conditioned blink reflexes could account for the differences in UR facilitation at the 100 and 150 msec SOAs seen in this experiment. Flaten (1993) and Flaten and Hugdahl (1991) observed increased eyeblink URs prior to observations of conditioned blink reflexes, and Weisz and McInerney (1990) found UR facilitation prior to the appearance of conditioned responses in nictitating membrane response conditioning in rabbits, a procedure similar to blink conditioning. They suggested that reflex facilitation was an index of a process underlying development of the conditioned eyeblink response. The mechanism behind this is probably facilitation of the motoneuron pool underlying the eyeblink UR (Guyton & Hall, 2006). That is, increased URs after CS presentation could be due to summation of sub-threshold activation induce by the CS, with suprathreshold activation induced by the US. Consequently, the impact of the US on the motoneurons is increased when the US is presented after the CS. An alternative explanation could be that the facilitated UR seen in the paired group was due to fear induced by the CS, or to a more specific process related to acquisition of the conditioned eyeblink reflex. There was no significant effect of Conditioning on ratings of the emotional valence of the CS. This does not lend support to the hypothesis that the CS induced fear in the paired group, which in turn facilitated eyeblink UR. However, due to the short duration of the ISI used in conditioning, a conditioned fear reaction to the CS would most likely be short-lived, as described by Davis et al. (1989). In that study, a 200 ms CS/US ISI supported a conditioned fear reaction, measured as UR facilitation, that was no longer observed at about 2000 ms. Thus, a shortlived fear reaction could theoretically be responsible for the present results, and this reaction may not have been picked up by the Self Assessment Manikin. There were significant correlations between CRs and the noise UR magnitudes at the 100, 150, and 1000 ms SOAs. There were no significant correlations between CRs and URs to the eyeblink-eliciting stimulus alone, and correlations between CR magnitudes and UR magnitudes at the 10, 30, and 50 ms SOAs were lower and not significant. Thus, there were correlations between UR facilitation and the CR at a time after CS onset when the CR was about to occur. Moreover, the correlations between UR facilitation and the CR increased from the first to the second half of trials. This is further support to the idea that conditioned facilitation of the eyeblink reflex indexes a process related to conditioned eyeblink responding. This supports Flaten's (1993)
21
and Weisz and McInerney's (1990) view that conditioned UR facilitation could be a result of a developing associative process that eventually produces the eyeblink CR. Overall this study supports Clark and Squire's (1998) view that simple delay conditioning may occur without awareness. Since reflex facilitation is observed just 100 ms after presentation of the CS, an automatic and pre-attentive process most likely mediates facilitation. Lovibond and Shanks (2002b) points to the many and substantial problems in assessing awareness by participants self-report. Reflex facilitation at short latencies may provide a method to bypass the question of awareness in classical conditioning. Clark and Squire (1998) argued that different systems mediate delay compared to trace conditioning, that is, trace conditioning is dependent on the hippocampus and, therefore, requires awareness. If that is the case, one would expect reflex potentiation at longer SOAs following trace conditioning compared to delay conditioning. This hypothesis should be tested in future studies. To conclude, the present study showed reflex facilitation as early as 100 ms after CS presentation in the paired compared to the unpaired group. This reflex facilitation was related to CR magnitude, indicating that the CR was mediated by an automatic and pre-attentive process. References Blumenthal, T.D., Elden, Å., Flaten, M.A., 2004. A comparison of several methods used to quantify prepulse inhibition of eyeblink responding. Psychophysiology 41, 326–332. Bradley, M., Lang, P.J., 1994. Measuring emotion — the self-assessment mannequin and the semantic differential. J. Behav. Ther. Exp. Psychiatry 25, 49–59. Clark, R.E., Squire, L.R., 1998. Classical conditioning and brain systems: the role of awareness. Science 280, 77–81. Clark, R.E., Squire, L.R., 2004. The importance of awareness for eyeblink conditioning is conditional: theoretical comment on Bellebaum and Daum (2004). Behav. Neurosci. 118, 1466–1468. Davis, M., Schlesinger, L.S., Sorenson, C.A., 1989. Temporal specificity of fear conditioning: Effects of different conditioned stimulus-unconditioned stimulus intervals on the fear-potentiated startle effect. J. Exp. Psychol. Anim. Behav. Process. 15, 295–310. Elden, A., Flaten, M.A., 2002. The relationship of automatic and controlled processing to prepulse inhibition. J. Psychophysiol. 16, 46–55. Elden, Å., Flaten, M.A., 2003. Similar effects of attention directed to acoustic and tactile stimuli an prepulse inhibition of acoustic startle. Scand. J. Psychol. 44, 361–370. Filion, D.L., Poje, A.B., 2003. Selective and nonselective attention effects on prepulse inhibition of startle: a comparison of task and no-task protocols. Biol. Psychol. 64, 283–296. Flaten, M.A., 1993. Startle reflex facilitation as a function of classical eyeblink conditioning in humans. Psychophysiology 30, 581–588. Flaten, M.A., Hugdahl, K., 1991. Does classical conditioning generate sensitization of the neural pathway of the conditional stimulus (CS+)? Psychobiology 19, 51–57. Flaten, M.A., Vaksdal, A., Hugdahl, K., 1989. An IBM-PC and commodore 64 microcomputer-based system for elicitation and recording of eyeblink reflexes. Biol. Psychol. 29, 291–298. Guyton, A.C., Hall, J.E. (Eds.), 2006. Textbook of Medical Physiology. Saunders, Philadelphia, PA, pp. 568–579. Lovibond, P.F., Shanks, D.R., 2002a. Automatic and eyeblink conditioning are closely related to contingency awareness: reply to Wiens and Öhman (2002) and Manns et al. (2002). J. Exp. Psychol. Anim. Behav. Process. 28, 38–42.
22
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
Lovibond, P.F., Shanks, D.R., 2002b. The role of awareness in Pavlovian conditioning: empirical evidence and theoretical implications. J. Exp. Psychol. Anim. Behav. Process. 28, 3–26. Manns, J.R., Clark, R.E., Squire, L.R., 2001. Single-cue delay eyeblink conditioning is unrelated to awareness. Cognit. Affect. Behav. Neurosci. 1, 192–198. O'Neill, S.T., Parrott, A.C., 1992. Stress and arousal in sedative and stimulant cigarette smokers. Psychopharmacology 107, 442–446. Smith, C.N., Clark, R.E., Manns, J.R., Squire, L.R., 2005. Acquisition of differential delay eyeblink classical conditioning is independent of awareness. Behav. Neurosci. 119.
Weisz, D.J., McInerney, J., 1990. An associative process maintains reflex facilitation of the unconditioned nictitating membrane response during the early stages of training. Behav. Neurosci. 104, 21–27. Wikgren, J., Ruusuvirta, T., Korhonen, T., 2002. Reflex facilitation during eyeblink conditioning and subsequent interpositus nucleus inactivation in the rabbit (Oryctolagus cuniculus). Behav. Neurosci. 116, 1052–1058.
International Journal of Psychophysiology 67 (2008) 17 – 22 www.elsevier.com/locate/ijpsycho
Conditioned facilitation of the unconditioned reflex after classical eyeblink conditioning Ole Åsli ⁎, Magne Arve Flaten Department of Psychology, University of Tromsø, N-9037 Tromsø, Norway Received 13 April 2007; received in revised form 18 September 2007; accepted 20 September 2007 Available online 4 October 2007
Abstract The present study (N = 40) investigated the time-course of conditioned facilitation of the unconditioned eyeblink reflex (UR). In a single-cue delay classical conditioning procedure, a tone conditioned stimulus (CS) signaled an airpuff unconditioned stimulus (airpuff US) to the eye. A paired group received 40 trials of CS/US presentations with an interstimulus interval (ISI) of 200 ms. An unpaired group received an equal number of explicitly unpaired presentations of the CS and US. Thereafter, eyeblink reflex facilitation was assessed by presenting 94 dB white noise (noise US) 10, 30, 50, 100, 150, and 1000 ms after CS onset. In the paired group, URs were significantly increased as early as 100 ms after CS onset compared to the unpaired group. This reflex facilitation was correlated with CR magnitude, indicating that conditioned facilitation of eyeblink URs indexes an early, automatic, preattentive stage in CR formation. © 2007 Elsevier B.V. All rights reserved. Keywords: Classical conditioning; Eyeblink conditioning; Conditioned facilitation; Reflex facilitation
1. Introduction Several studies have found that unconditioned reflexes (URs), elicited in the presence of a conditioned stimulus (CS), are facilitated compared to URs elicited in the absence of the CS (Flaten 1993; Flaten and Hugdahl 1991; Weisz and McInerney 1990; Wikgren et al. 2002). The present study investigated the time-course of conditioned facilitation of URs. It has been hypothesized that conditioned facilitation of the UR indexes an early stage of eyeblink CR formation (Flaten 1993; Weisz and McInerney 1990). Flaten (1993) used an airpuff to the eye as the US, and found eyeblink UR facilitation to a noise US after only four CS/US pairings, prior to the occurrence of eyeblink conditioned responses (CRs). Similar findings were made by Weisz and McInerney (1990) and Wikgren et al. (2002). It was argued that conditioned UR facilitation was the result of a developing associative process that would eventually produce the eyeblink CR. The process underlying UR facilitation could be CS-induced facilitation of the motoneuron pool in the facial ⁎ Corresponding author. Tel.: +47 776 46778; fax: +47 776 45610. E-mail address: [email protected] (O. Åsli). 0167-8760/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpsycho.2007.09.003
nerve that is responsible for the eyeblink (e.g., Guyton and Hall, 2006). After CS presentation, it is hypothesized that the motoneuron pool is gradually more facilitated until a CR is elicited or not. Consequently, when a reflex-eliciting stimulus is presented to an already facilitated motoneuron pool, the magnitude of the reflex increases. The degree of increase of the reflex would depend on the number of facilitated neurons and the amount of facilitation or depolarization. There is disagreement on the role of voluntary responding and awareness of the CS/US contingency in eyeblink classical conditioning. (Clark and Squire 2004; Manns et al. 2001; Smith et al. 2005) argued that single-cue delay classical conditioning is independent of awareness of the CS/US contingency. However, their measure of awareness was criticized by Lovibond and Shanks (2002a) for underestimating awareness of the CS / US contingency. Lovibond and Shanks (2002b) argued that the claim that automatic, non-conscious processes played a role in human conditioning had not been proven. Rather, a conscious, propositional system could explain the development of the CR. One central point in this debate is the claim by Lovibond and Shanks (2002b) that almost every study of the role of consciousness in classical conditioning used inadequate measures of
18
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
awareness, and that most measures tend to underestimate awareness, or that the experimental procedures used may lead to underestimated awareness. The present study bypassed the problem of awareness and voluntary responding by investigating CR and UR facilitation at points in time after CS onset that were too short for conscious processes to play a role. The participants in the paired group first received 40 trials of eyeblink conditioning with a short interstimulus interval (ISI) of 200 ms. The short ISI was used to reduce the role of voluntary responding in the CR data. In the unpaired group, an equal number of CSs and USs were presented in an explicitly unpaired manner. Thereafter, UR facilitation was tested by presenting eyeblink reflex-eliciting noise immediately after CS presentation. The stimulus onset asynchronies (SOAs) between CS onset and noise onset were chosen to investigate the latency and development of the process of UR facilitation. A positive correlation between CR amplitude and UR facilitation was expected if the latter could be related to a process underlying CR production. Evaluations of CS and US valence and arousal were obtained to investigate whether emotional processes could be related to UR facilitation. 2. Materials and methods 2.1. Subjects Forty subjects (23 females and 17 males, age range 19–37, mean age 25.9 years) participated in the study. The participants were randomly assigned to one of two groups (paired group or unpaired group), each containing 20 participants. All participants were in good health and did not report any previous serious disease or injury. All participants had auditory thresholds of 30 dB or less at 1000 Hz in both ears. The project was approved by the Regional Committee of Medical Research Ethics in Health Region IV in Norway (Project 29/2000). The participants were instructed to refrain from drinking caffeinated beverages and use nicotine-containing substances for 3 hours prior to the start of the study. Written informed consent was obtained from all participants, who where paid 2 lottery tickets (equivalent to 50 NOK) for their participation. 2.2. Apparatus and stimuli The experiment took place in an electrically and acoustically shielded chamber (C. A. Tegnèr), where the temperature was kept at about 20 °C. Hearing was tested with a Grason-Stadler 17 audiometer. A Bruel and Kjær 2235 Sound Level Precision Meter measured intensity of auditory stimuli. Programs written by the last author in ASYST 3.1 and run on an ALR 486 PC, controlled presentation of experimental stimuli and data acquisition. There were two different unconditioned stimuli (US) in the experiment. The airpuff US was paired with the conditioned stimuli (CS). The noise US was deployed to elicit eyeblink UR and was not paired with a CS. The noise US had an intensity of 94 dB, instantaneous rise time and a duration of 50 ms. The CSs were 1000-Hz tones with intensity of 60 dB, and a rise time of 20 ms. The duration of the prepulses were the same as the
stimulus onset asynchrony (SOA) between the tone prepulse and the eyeblink-eliciting noise. The stimuli were delivered through Sennheiser HD 250 headphones. The airpuff US was directed to the corner of the right eye. The airpuffs had an intensity of 20 kPa, as indicated at the pressure regulator on the air-pressure bottle, and a duration of 50 ms. The airpuffs were delivered by apparatus described in Flaten et al. (1989). Eyeblink electromyographic (EMG) responses were recorded from the right orbiculari oculi with two Ag/AgCl Sensor Medics miniature electrodes (2 mm diameter) filled with Ultra Phonic conductivity gel. Inter-electrode distance was 1.5– 2.5 cm. The ground electrode was placed centrally on the forehead. The EMG signal was amplified by a factor of 60,000 and filtered (passing 90–250 Hz) by a Coulbourn S75-01 bioamplifier. The signal was rectified and integrated with a Coulbourn S76-01 contour-following integrator with a 10-ms time constant, and the output sent to the PC via a Keithley interface. Sampling on each trial began 200 ms prior to onset of the first stimulus and continued for 200 ms after onset of the US. The sampling rate was 10 Hz prior to onset of the first stimulus and 1000 Hz after stimulus onset. Emotional valence and arousal elicited by the CS and airpuff US was recorded with the Self Assessment Manikin (SAM) (Bradley and Lang 1994). General stress and arousal was recorded with the O'Neill and Parrott (1992) stress and arousal scale. 2.3. Procedure After arrival at the laboratory the subjects sat down in a desk chair and read and filled in the Informed Consent Form. Level of hearing was tested before the participants were lead into the experimental chamber and seated in a reclining chair. The subjects were informed of the general purpose of the study, about the stimuli and procedure, and that they could withdraw from the study without giving any reason at any time. The skin below the participants' right eye was cleaned with a swab containing alcohol and pumice, and the electrodes for measurement of the eyeblink electromyography (EMG) were attached. The headgear containing the airpuff-presentation equipment and the headphones were attached, and then the subjects filled out the O'Neill and Parrott (1992) stress and arousal scale. The door to the experimental chamber was closed, and after a brief pause the experimental procedure was in effect. The experiment consisted of two phases (see Fig. 1). In the conditioning phase, the paired group received 40 trials of singlecue classical conditioning with the 1000 Hz tone as the CS and the airpuff to the eye as the US. The interstimulus interval (ISI) from CS onset to US onset was 200 ms. The unpaired group received 40 explicitly unpaired presentations of the CS and US. In the noise UR phase, UR facilitation to the noise US was investigated by presenting the CS at SOAs of 10, 30, 50, 100, 150, and 1000 milliseconds relative to onset of the eyeblink eliciting noise. Each SOA was presented four times. The noise US was also presented four times alone, so a total of 28 trials were presented. The SOAs were presented in semi-random
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
19
onset. Thus, alpha responses could not be scored mistakenly as CRs. To ascertain this, a separate analysis of alpha responses was performed that confirmed that there was no difference between the paired group and the unpaired group in alpha responses during conditioning. 2.5. Design and statistics The design for the conditioning phase was a 2 Conditioning (paired, unpaired) × 8 Trial blocks mixed design where the first factor was treated as a between-subjects factor and the last factor was treated as a within-subjects factor. The design for the noise UR phase of the experiment was a 2 Conditioning (paired, unpaired) × 7 SOA (noise alone, 10, 30, 50, 100, 150, and 1000 ms) mixed design with the first factor treated as a between-subjects factor and the last factor treated as a within-subjects factor. The data were analysed with repeated measures analysis of variance (Statistica 7.0). Theoretically interesting contrasts were examined by contrast analyses. 3. Results 3.1. Subjective data There were no significant main effects or interactions in the ratings of CS and US valence and arousal. 3.2. Conditioned responses Fig. 1. Schematic presentation of the procedure.
order in two blocks of 12 trials each. The airpuff US was presented 400 ms after CS onset at 20 trials in the paired group in this phase to avoid extinction. The airpuff US was not presented on trials where the SOA was 1000 ms. At the end of the experiment the subjects filled in the O'Neill and Parrott posttest, and two SAM forms for recording of affective reactions to the CS and airpuff US.
There was a significant main effect of Conditioning (F(1,38) = 13.23, p = .0008), due to significantly larger conditioned responses in the paired group compared to the unpaired group. There was a significant interaction of Conditioning by Trial block (F(7,266) = 3.02, p = .004), and follow-up tests showed significantly stronger CRs in the paired group compared to the unpaired group in Trial blocks 3, 4, 5, 6, 7 and 8 (all ps b .006) (Fig. 3.). The largest difference in CR magnitude between the groups was 5.9 μV in Trial block 6.
2.4. Response scoring and data reduction Noise unconditioned eyeblink reflexes were scored as the difference between the mean of a 200 ms baseline recorded prior to the first stimulus presented in that trial, and the maximum EMG level in the interval from 20 to 120 ms after onset of the noise. Noise URs were also expressed as proportion of difference from control: Average responding to each SOA was calculated across the entire session. Average responding to noise US alone (control) was subtracted from the average responses to each SOA and this difference was divided by the average responding to the noise US alone (Blumenthal et al., 2004). Conditioned eyeblink reflexes (CRs) were scored as the difference between the mean of a 200 ms pre-stimulus baseline and the maximum EMG level in the interval from 130 to 200 ms after CS onset. This window was chosen since alpha responses, i.e., unconditioned responses occasionally occurring after CS onset, are usually terminated between 100 and 120 ms after CS
Fig. 2. Left panel: Conditioned eyeblink responses in the paired and unpaired groups in the conditioning phase. Error bars represent 1 standard error of the mean. Right panel: Noise UR magnitudes across SOAs. S2 denotes noise alone trials.
20
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
Fig. 3. Grand average of responses to the noise US alone and to the noise US presented 100 ms after the CS, in the paired and unpaired groups.
3.3. Noise unconditioned response magnitude There was no difference between the paired and the unpaired group in noise UR magnitude to noise alone (F b 1, ns). The main effect of SOA was significant (F(6, 228) = 31.08, p = .00001). There was also a significant interaction of Conditioning by SOA (F(6, 228) = 11.80, p = .000001) and follow-up tests showed significantly more facilitation of noise URs elicited 100 and 150 ms after CS onset in the paired group compared to the unpaired group ( p = .004 and p = .0002, respectively) (Fig. 2). The greatest difference in noise UR magnitude between the paired and the unpaired group was at 150 ms, where there was a difference of about 14.9 μV. Fig. 3 shows the mean EMG for each millisecond for all acceptable trials for the noise-elicited Table 1 Correlations between CRs in the first 20 trials (First) and the last 20 trials (Last) in the conditioning phase, and noise-elicited eyeblink URs across the SOAs in the noise UR phase
S2 10 30 50 100 150 1000 S2 denotes noise alone trials. a p b .05.
First
Last
0.11 0.17 0.16 0.17 0.41 a 0.46 a 0.29
0.24 0.27 0.24 0.30 0.56 a 0.55 a 0.37 a
UR to the noise alone, and for noise-elicited URs to the noise presented 100 ms after CS onset, i.e. at the shortest SOA where significant reflex facilitation was observed. Noise-elicited URs expressed as proportion of difference from control revealed similar results as UR magnitude analyses. 3.4. Correlations between CR magnitude and UR magnitude There were significant positive correlations between CRs in the second half of the trials in the Conditioning phase, and noise URs at the 100, 150, and 1000 ms SOAs in the Noise UR phase (Table 1). 4. Discussion There was a significant interaction of Conditioning by SOA in the eyeblink UR data, due to significantly increased URs at the 100 and 150 ms SOAs in the paired group compared to the unpaired group. This short latency of facilitated UR indicates that the mediating process is automatic and pre-attentive. As attentional modulation of URs has been seen at short SOAs (Elden and Flaten 2002, 2003; Filion and Poje 2003), such an interpretation is worth mentioning. Elden and Flaten (2002) compared task and no-task conditions presented in separate blocks, and found significantly greater inhibition of the UR in the task than no-task condition at SOAs between 30 and 420 ms (Experiment 2). This led the researchers to conclude that preparatory attention, i.e., attention to a stimulus about to occur,
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
inhibited subsequently elicited UR. Preparatory attention to the CS could have been induced in the participants in the present study. However, preparatory attention cannot explain the present results, since attention to the CS in the paired group would have inhibited the UR, i.e., the opposite of the facilitated the UR seen in the present experiment. The UR facilitation seen in the paired group was most likely not CRs mistakenly scored as URs to the noise. The largest difference in CRs between the paired and unpaired groups in the conditioning phase was 5.9 μV in trial block six, a relatively small difference compared to the difference of 14.9 μV in UR facilitation in the noise UR phase. Thus, it is unlikely that conditioned blink reflexes could account for the differences in UR facilitation at the 100 and 150 msec SOAs seen in this experiment. Flaten (1993) and Flaten and Hugdahl (1991) observed increased eyeblink URs prior to observations of conditioned blink reflexes, and Weisz and McInerney (1990) found UR facilitation prior to the appearance of conditioned responses in nictitating membrane response conditioning in rabbits, a procedure similar to blink conditioning. They suggested that reflex facilitation was an index of a process underlying development of the conditioned eyeblink response. The mechanism behind this is probably facilitation of the motoneuron pool underlying the eyeblink UR (Guyton & Hall, 2006). That is, increased URs after CS presentation could be due to summation of sub-threshold activation induce by the CS, with suprathreshold activation induced by the US. Consequently, the impact of the US on the motoneurons is increased when the US is presented after the CS. An alternative explanation could be that the facilitated UR seen in the paired group was due to fear induced by the CS, or to a more specific process related to acquisition of the conditioned eyeblink reflex. There was no significant effect of Conditioning on ratings of the emotional valence of the CS. This does not lend support to the hypothesis that the CS induced fear in the paired group, which in turn facilitated eyeblink UR. However, due to the short duration of the ISI used in conditioning, a conditioned fear reaction to the CS would most likely be short-lived, as described by Davis et al. (1989). In that study, a 200 ms CS/US ISI supported a conditioned fear reaction, measured as UR facilitation, that was no longer observed at about 2000 ms. Thus, a shortlived fear reaction could theoretically be responsible for the present results, and this reaction may not have been picked up by the Self Assessment Manikin. There were significant correlations between CRs and the noise UR magnitudes at the 100, 150, and 1000 ms SOAs. There were no significant correlations between CRs and URs to the eyeblink-eliciting stimulus alone, and correlations between CR magnitudes and UR magnitudes at the 10, 30, and 50 ms SOAs were lower and not significant. Thus, there were correlations between UR facilitation and the CR at a time after CS onset when the CR was about to occur. Moreover, the correlations between UR facilitation and the CR increased from the first to the second half of trials. This is further support to the idea that conditioned facilitation of the eyeblink reflex indexes a process related to conditioned eyeblink responding. This supports Flaten's (1993)
21
and Weisz and McInerney's (1990) view that conditioned UR facilitation could be a result of a developing associative process that eventually produces the eyeblink CR. Overall this study supports Clark and Squire's (1998) view that simple delay conditioning may occur without awareness. Since reflex facilitation is observed just 100 ms after presentation of the CS, an automatic and pre-attentive process most likely mediates facilitation. Lovibond and Shanks (2002b) points to the many and substantial problems in assessing awareness by participants self-report. Reflex facilitation at short latencies may provide a method to bypass the question of awareness in classical conditioning. Clark and Squire (1998) argued that different systems mediate delay compared to trace conditioning, that is, trace conditioning is dependent on the hippocampus and, therefore, requires awareness. If that is the case, one would expect reflex potentiation at longer SOAs following trace conditioning compared to delay conditioning. This hypothesis should be tested in future studies. To conclude, the present study showed reflex facilitation as early as 100 ms after CS presentation in the paired compared to the unpaired group. This reflex facilitation was related to CR magnitude, indicating that the CR was mediated by an automatic and pre-attentive process. References Blumenthal, T.D., Elden, Å., Flaten, M.A., 2004. A comparison of several methods used to quantify prepulse inhibition of eyeblink responding. Psychophysiology 41, 326–332. Bradley, M., Lang, P.J., 1994. Measuring emotion — the self-assessment mannequin and the semantic differential. J. Behav. Ther. Exp. Psychiatry 25, 49–59. Clark, R.E., Squire, L.R., 1998. Classical conditioning and brain systems: the role of awareness. Science 280, 77–81. Clark, R.E., Squire, L.R., 2004. The importance of awareness for eyeblink conditioning is conditional: theoretical comment on Bellebaum and Daum (2004). Behav. Neurosci. 118, 1466–1468. Davis, M., Schlesinger, L.S., Sorenson, C.A., 1989. Temporal specificity of fear conditioning: Effects of different conditioned stimulus-unconditioned stimulus intervals on the fear-potentiated startle effect. J. Exp. Psychol. Anim. Behav. Process. 15, 295–310. Elden, A., Flaten, M.A., 2002. The relationship of automatic and controlled processing to prepulse inhibition. J. Psychophysiol. 16, 46–55. Elden, Å., Flaten, M.A., 2003. Similar effects of attention directed to acoustic and tactile stimuli an prepulse inhibition of acoustic startle. Scand. J. Psychol. 44, 361–370. Filion, D.L., Poje, A.B., 2003. Selective and nonselective attention effects on prepulse inhibition of startle: a comparison of task and no-task protocols. Biol. Psychol. 64, 283–296. Flaten, M.A., 1993. Startle reflex facilitation as a function of classical eyeblink conditioning in humans. Psychophysiology 30, 581–588. Flaten, M.A., Hugdahl, K., 1991. Does classical conditioning generate sensitization of the neural pathway of the conditional stimulus (CS+)? Psychobiology 19, 51–57. Flaten, M.A., Vaksdal, A., Hugdahl, K., 1989. An IBM-PC and commodore 64 microcomputer-based system for elicitation and recording of eyeblink reflexes. Biol. Psychol. 29, 291–298. Guyton, A.C., Hall, J.E. (Eds.), 2006. Textbook of Medical Physiology. Saunders, Philadelphia, PA, pp. 568–579. Lovibond, P.F., Shanks, D.R., 2002a. Automatic and eyeblink conditioning are closely related to contingency awareness: reply to Wiens and Öhman (2002) and Manns et al. (2002). J. Exp. Psychol. Anim. Behav. Process. 28, 38–42.
22
O. Åsli, M.A. Flaten / International Journal of Psychophysiology 67 (2008) 17–22
Lovibond, P.F., Shanks, D.R., 2002b. The role of awareness in Pavlovian conditioning: empirical evidence and theoretical implications. J. Exp. Psychol. Anim. Behav. Process. 28, 3–26. Manns, J.R., Clark, R.E., Squire, L.R., 2001. Single-cue delay eyeblink conditioning is unrelated to awareness. Cognit. Affect. Behav. Neurosci. 1, 192–198. O'Neill, S.T., Parrott, A.C., 1992. Stress and arousal in sedative and stimulant cigarette smokers. Psychopharmacology 107, 442–446. Smith, C.N., Clark, R.E., Manns, J.R., Squire, L.R., 2005. Acquisition of differential delay eyeblink classical conditioning is independent of awareness. Behav. Neurosci. 119.
Weisz, D.J., McInerney, J., 1990. An associative process maintains reflex facilitation of the unconditioned nictitating membrane response during the early stages of training. Behav. Neurosci. 104, 21–27. Wikgren, J., Ruusuvirta, T., Korhonen, T., 2002. Reflex facilitation during eyeblink conditioning and subsequent interpositus nucleus inactivation in the rabbit (Oryctolagus cuniculus). Behav. Neurosci. 116, 1052–1058.