Revisiting the James versus Cannon debate onemotion: startle and autonomic modulation in

patients with spinal cord injuries

Pilar Cobos a,*, Marıa Sanchez b, Carmen Garcıa c,Marıa Nieves Vera b, Jaime Vila b

a Departamento de Personalidad, Evaluacion y Tratamiento Psicologico, Facultad de Psicologıa, Universidad

de Malaga, Malaga, Spainb Departamento de Personalidad, Evaluacion y Tratamiento Psicologıca, Facultad de Psicologı, Universidad

de Granada, Granada, Spainc Departamento de Tecnologıa Electronica de la E. T. Superior de Telecomunicacion, Universidad de Malaga,

Malaga, Spain

Abstract

James’ hypothesis that impaired peripheral physiology in patients with spinal cord injuries

(SCI) impairs emotional processing, as manifested in the modulation of physiological

responses and in the subjective component of emotions, was examined in the present study.

A pilot study confirmed the utility of Lang’s picture viewing paradigm in a group of 78

students using the Spanish norms of the International Affective Picture System. In the main

study, 19 patients with SCI and 19 well controls matched for sex, age and education were

examined. Results showed: (1) no differences between SCI and control participants in the

valence and arousal ratings of the pictures; (2) similar heart rate modulation in both groups,

i.e. the unpleasant pictures produced greater deceleration than the pleasant ones; and (3) no

decrease in emotional experience in the SCI group compared with the control group. The

implications of the results for the James versus Cannon controversy on the theory of emotions

are discussed.

# 2002 Elsevier Science B.V. All rights reserved.

Keywords: Emotions; Spinal cord injuries; Startle; Heart rate; Skin conductance

* Corresponding author

E-mail addresses: mpcobos@uma.es (P. Cobos), jvila@ugr.es (J. Vila).

Biological Psychology 61 (2002) 251�/269

www.elsevier.com/locate/biopsycho

0301-0511/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.

PII: S 0 3 0 1 - 0 5 1 1 ( 0 2 ) 0 0 0 6 1 - 3

1. Introduction

Psychologists have repeatedly debated the implication of peripheral bodily

responses in the experience and expression of emotions. The old debate between

supporters of the James (James, 1884, 1890, 1894) and Cannon (Cannon 1914, 1927,

1929) theories established the coordinates from which later theories have evolved:

peripheralism versus centralism, bodily responses versus cognition, and specificityversus dimensionality. These dichotomous perspectives have remained present

throughout the history of psychology, the emphasis varying according to the

dominant paradigm*/behaviorist versus cognitive*/with few attempts being made

at integration.

One of these attempts is that proposed by Lang in the context of his

bioinformational and hierarchical theory of emotions (Birbaumer and Ohman,

1993; Bradley, 2000; Lang, 1979, 1994, 1995; Lang et al., 1997, 2000). Basically,

Lang’s model understands human emotions as action dispositions that aredetermined by the activation of specific brain circuits. These circuits are located in

deep cortical and subcortical structures, being closely related to two primary

motivational systems: the appetitive and the defensive. When these circuits are

activated, the physiological, behavioral and subjective manifestations of emotions

arise, exerting a modulatory effect on the brain’s other processing operations,

including the potentiation or inhibition of simple exteroceptive reflexes. Lang’s

model assumes that the functional architecture of emotion is hierarchically organized

along elements that are both specific and dimensional, physiological and cognitive,and peripheral and central, providing a new theoretical context in which the old

James�/Cannon debate can be tested. Research data supporting Lang’s model have

mainly been obtained using two paradigms: mental imagery and picture viewing

(Lang, 1985, 1995).

A line of research specifically designed to investigate the James�/Cannon

controversy on emotion, yet to be studied under Lang’s paradigms, is that examining

patients with spinal cord injuries (SCI). Ever since the seminal paper by James

(1884), patients with different degrees of reduction in their peripheral feedback havebeen considered ideal participants for testing these theories. Studies on emotions in

participants with spinal cord lesions have almost exclusively used self-report

measures (Dana, 1921; Hohmann, 1966; Richards et al., 1982; Lowe and Carroll,

1985; Chwalisz et al., 1988; Bermond et al., 1991), with their associated

methodological biases, including investigator bias due to social desirability and

demand characteristics. Indeed, the results of these studies do not all point in the

same direction and have been criticized for methodological flaws (Chwalisz et al.,

1988; Trieschmann, 1980; Tucker, 1980; Reisenzein, 1983; Richards et al., 1982), forabsence of a complete lesion (Fehr and Stern, 1970), or for failures to control for

medication and other relevant variables (Chwalisz et al., 1988).

The aim of the present study was to examine the emotional modulation of

psychophysiological responses in patients with SCI using Lang’s picture viewing

paradigm. This paradigm combines the visualization of affective pictures with the

presentation of a brief burst of white noise to elicit the startle reflex. In addition to

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269252

the subjective assessments, this paradigm can provide uncontaminated evidence

against or in favor of the role of peripheral bodily responses in the processing of

emotional experience and expression, helping to clarify the old controversies. Lang’s

paradigm was recently used to study completely paralyzed patients with amiotrophic

lateral sclerosis (ALS), and yielded evidence against the idea of impaired emotional

processing in patients with severely impaired peripheral physiology (Birbaumer,

2001). ALS is a severe neurological disease that progressively destroys motorneurons, leading to a total loss of voluntary muscular responses. Communication

with this type of completely paralyzed locked-in patients is possible using techniques

such as the thought-translation-device developed by Birbaumer et al. (1999), and

based on the recording and self-regulation of slow cortical potentials, or the brain�/

computer Interface technology developed by Wolpaw et al. (2000)), and based on the

recording and self-regulation of spontaneous EEG activity (for a description of

alternative brain�/computer interfaces, see Kubler et al., 2001). Using the thought-

translation-device, Birbaumer has reported normal reactivity of ALS patients in thepicture viewing paradigm as indicated by EMG, heart rate, and subjective ratings of

affective pictures.

The startle reflex is a chained series of rapid flexor movements prompted by any

abrupt sensory stimulation. In humans, the first and most reliable component of the

reflex is a rapid eyeblink. Eyeblink startle has been considered a defensive reflex that

serves a protective function, helping to avoid organ injury (Pavlov, 1927; Landis &

Hunt, 1939) and acting as a behavioral interrupt that clears the processor to deal

with a possible threat (Graham, 1979; Lang, 1995). In the picture viewing paradigm,eyeblink magnitude is consistently potentiated when participants view unpleasant

pictures and inhibited when they view pleasant pictures (Lang, 1995; Bradley, 2000).

This emotional modulation of the startle reflex is explained by Lang and colleagues

as a motivational priming effect: unconditioned reflexes are primed that are linked to

appetitive or defensive motivational systems. Thus, activation of the defensive

motivational system by viewing unpleasant pictures should prime defensive reflexes,

whereas activation of the appetitive motivational system by viewing pleasant pictures

should lead to inhibition of defensive responding (Bradley, 2000).In the context of the picture viewing paradigm, other physiological responses have

also shown consistent covariations with subjective ratings of picture-elicited valence

and arousal: heart rate, skin conductance, facial EMG, and event-related potentials.

Heart rate and skin conductance are two autonomic physiological responses of

special interest for our research, because they are presumably impaired in SCI

patients. Heart rate shows a consistent pattern of decelerative changes during the 6 s

of picture viewing: an initial deceleration during seconds 1�/2, a subsequent

acceleration during seconds 3�/4, and a final greater deceleration during seconds5�/6. The decelerative pattern is significantly greater when viewing unpleasant

pictures versus pleasant ones, while the neutral pictures elicit an intermediate

pattern. Skin conductance also shows a consistent pattern, characterized by a slow

increase after the first 2 s that does not recover within the 6 s picture presentation.

The increase is significantly greater when viewing both pleasant and unpleasant

pictures than when viewing neutral pictures.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 253

The present study tested the specific hypothesis that impaired peripheral

physiology in patients with SCI impairs emotional processing, as manifested in

modulation of physiological responses (eyeblink startle, heart rate and skin

conductance) as well as in the subjective component of emotions. Regarding the

subjective component, the study by Hohmann (1966) was taken as the classic

methodological reference for the study of emotions in people who have suffered

spinal cord lesions. In addition to the emotions and affects studied by Hohmann(fear, anger, sadness, and sentimentalism), our study included scales to evaluate

changes in other emotions (joy and love). The study was preceded by a pilot study to

test the picture viewing task in students, using the Spanish norms of the international

affective picture system (Molto et al., 1999; Vila et al., 2001) and the data acquisition

and experimental control software developed in our lab in the context of the present

research (Garcıa et al., 1997).

2. Pilot study

2.1. Participants

A total of 78 undergraduate volunteer students (66 females, 12 males) participated

in the study (age range�/20�/23 years). None of the participants reported being

under pharmacological or psychiatric treatment and none had cardiac disorders or

auditory or visual deficiencies.

2.2. Design

A within-subjects design was employed, with valence as a repeated measure factor

with three levels: pleasant, neutral, and unpleasant.

2.3. Task

Thirty pictures, consisting of 10 pleasant, 10 neutral, and 10 unpleasant pictures,were selected from the IAPS on the basis of their valence ratings (scale 1�/9)

according to Spanish norms (Molto et al., 1999). Pleasant pictures were within the

7.86�/8.3 range, neutral pictures within the 4.06�/4.97 range, and unpleasant pictures

within the 1.38�/2.14 range. Participants saw the 30 pictures in a random order with

the following limitations: each block of 10 pictures included at least 2 pictures of

each valence category, and no more than 3 pictures belonging to the same category

were presented consecutively. Each picture was presented for 6 s with an inter-trial

interval of 24 s. The acoustic startle probe was randomly delivered between seconds 2and 4 during half of the pictures in each category. The task was preceded by a 5-min

non-stimulation period and finished 1 min after the last picture. IAPS identification

number of the pictures used are: (pleasant) 2080, 2360, 2070, 1710, 2170, 2540, 8490,

7330, 1750, 1920; (neutral) 6150, 5920, 1310, 9270, 2230, 1390, 7700, 1120, 7560,

3250; (unpleasant) 3180, 3010, 2710, 3230, 9400, 9420, 3170, 9410, 2800, 3000.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269254

2.4. Apparatus

The following apparatus was used: (1) a Letica polygraph (Model 4006) to record

integrated EMG through the 905 EMG amplifier. Recording was done using a band

filter of 10�/1000 Hz. Rectification and contour-following-integration was set at 150

ms time constant, with a sensitivity of between 100 and 200 mV/full scale; (2) a

Reflecta (model MC 150) slide projector to display the slides; and (3) a Letica (modelLE 150) auditory stimulator to generate white noise. The sounds were presented

through earphones (model telephonics) and the intensity was calibrated with a Bruel

and Kjaer sonometer (model 2235) using an artificial ear (model 4153).

Data acquisition and stimulus control were monitored by the BioLab program,

software specifically developed by the Department of Telecommunication Engineer-

ing of Malaga University for this research (see Cobos, 1999). It uses a 12-bit PCL-

812PG Advantech card on a PC Pentium and MS-DOS together with some

LabWindows libraries. The physiological variable was recorded at a voltage rangeof �/5 to �/5 V and a sampling rate of 1000 Hz, starting 1 s before picture onset and

finishing 3 s after picture offset.

2.5. Dependent variables

The eyeblink component of the startle response was measured by recording

integrated EMG activity from the orbicularis oculi muscle beneath the left eye, usingsmall Sensormedic electrodes filled with electrolyte paste. Eyeblink magnitude was

scored as the maximum increase in microvolts occurring between 20 and 120 ms after

noise onset. In addition, valence and arousal ratings for each picture were obtained

by using the self-assessment-manikin (SAM) scale (Lang et al., 1988) to evaluate the

pleasantness and arousal of each picture after the task. This scale ranges from 1 to 9

points.

2.6. Procedure

Each participant underwent a single session at the laboratory. After completing

the informed consent form, the participant reclined in a comfortable chair and the

EMG sensors were attached. The participant was then instructed to view each

picture for the entire time it was on the screen. In addition, he/she was told that

occasional noises would be heard through the headphones and could be ignored.

After checking the physiological recording, the beginning of the test was signaled by

turning down the lighting in the participant’s room to a pre-established subduedlevel. After the picture series was finished, the sensors were removed and the same

picture series was presented again in the same order, with an inter-picture interval of

12 s, in order to evaluate the pleasantness and arousal of each picture using the SAM

scale. Finally, the participant was debriefed and given the academic credit for his/her

participation.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 255

2.7. Data analysis

Analysis of variance for repeated measures was applied to the average eyeblink

magnitude in the five probe trials within each valence category. The Greenhouse

Geisser epsilon correction was applied for the adjustment of the degrees of freedom.

Results are reported with the original degrees of freedom and the corrected P -values.

Post-hoc comparisons were performed using the Tukey test. The subjective variableswere analyzed using the Friedman test (x2).

2.8. Results

2.8.1. Subjective ratings

Average valence ratings were: pleasant (7.74, S.D.�/1.33), neutral (4.32, S.D.�/

1.34), and unpleasant (1.78, S.D.�/1.08). Results of the Friedman test showed

significant differences between the valence ratings (x2(2, 78)�/155.51, PB/ 0.0001).

The differences were significant between the three valence categories. Average

arousal ratings were: pleasant (3.34, S.D.�/1.55), neutral (3.35, S.D.�/1.14), and

unpleasant (6.83, S.D.�/1.25). Results of the Friedman test also showed significant

differences between the arousal ratings (x2(2, 78)�/117.62, PB/ 0.0001). In this case,

the difference were significant only between the unpleasant and the other twocategories; no significant differences was observed between the neutral and pleasant

categories.

2.8.2. Eyeblink

Fig. 1 plots the eyeblink magnitude as a function of the affective valence of the

pictures. Results of the ANOVA showed a significant effect of valence (F (2, 154)�/

7.85, PB/ 0.001), with a significant linear trend (F(1, 77)�/15.6, PB/ 0.001). Post hoc

comparisons showed significant differences between the pleasant and neutral (PB/

Fig. 1. Eyeblink magnitude to the startle noise during pleasant, neutral and unpleasant picture viewing in

the pilot study. Error bars are one standard error.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269256

0.05), and between the pleasant and unpleasant (PB/ 0.001) categories; no significant

difference was observed between the neutral and unpleasant categories.

2.9. Discussion

The startle results reproduced the expected valence effect reported by Lang and

colleagues. The magnitude of the startle response to the probe was greater when

participants were viewing the unpleasant pictures than when viewing the pleasant

ones, with the neutral pictures producing an intermediate magnitude. Results of thesubjective ratings confirmed the a priori selection of the pictures based on the

Spanish norms of the International Affective Picture System. As expected, the

participants rated the pleasant pictures as significantly more pleasant than the

neutral and unpleasant ones, and the neutral pictures as significantly less unpleasant

than the unpleasant ones. The arousal ratings showed significant differences between

the unpleasant and neutral categories and between the unpleasant and pleasant ones.

However, the arousal ratings of the pleasant pictures were not significantly different

from those of the neutral ones. The latter result is not uncommon when the selectionof pictures is based only on valence ratings, because both American and Spanish

students consistently rate unpleasant materials as more arousing than pleasant ones

(Bradley, 2000; Molto et al., 1999; Vila et al., 2001). In general, the results of the

pilot study were considered adequate to extend the same methodology to test our

main hypothesis, which addressed the emotional modulation of eyeblink startle and

other physiological responses in SCI patients.

3. Main study

3.1. Participants

The study included 19 participants with SCI, 7 women (aged 21�/53 years) and 12

men (aged 24�/57 years); and 19 control participants, 7 women and 12 men, matched

for age and education. The physical and health status of the patients was good,

within the obvious limits imposed by their lesions, and they had no psychiatric

problems or dependence on drugs or alcohol. None of the participants had received

any psychological or psychiatric therapy after their lesion, the duration of which

ranged from 22 months to 28 years. The characteristics of the lesions and the age,sex, duration of the lesion, marital status, and education of the patients are shown in

Table 1. The inclusion criterion for the patient group was a minimum 18-month

history of complete or incomplete traumatic or surgical spinal cord injury. Exclusion

criteria were: any degree of mental handicap, severe visual or auditory deficit, and/or

higher nervous system involvement.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 257

3.2. Design

The physiological variables were compared between injured and noninjured

participants using a factorial design with two independent groups and several

repeated measures factors, as specified below. The subjective variables were

compared as follows: first, by using a general two group design (patients and

controls); second, by dividing the 19 SCI patients into three groups, according to the

level of their lesion: High lesions (the six with highest lesions), Middle lesions (theseven with lesions at an intermediate level), and Low lesions (the six with the lowest

lesions); third, by dividing the 19 SCI patients into two groups according to the

extent of their lesion: complete lesion (12 patients with complete section of the spinal

cord) and incomplete lesion (7 patients with a partial section of the spinal cord); and

fourth, by dividing the patients into two groups according to the duration of the

lesion: 10 patients with duration B/10 years and 9 patients with duration E/10 years.

3.3. Task

The task was the same as in the pilot study.

Table 1

Lesion and socio-demographic characteristics of the SCI patients

Level and extent of the injury Age (years/month) Years since lesion Education Marital status

Men

C1-2a 24.05 7 Se S

C4a 39.01 21 Su S

T4a 40.08 26 Se S/P

T6a,c 48.04 23 Su S/P

T6a 30 1.11 Se M

T6b 24 4 Pri S

T7a 31.10 16 Pri S/P

T8b,c 33.04 10 Se S/P

T8-9b 40.06 5 Pri M

T10a,c 52 27 Pri M

T11a 57.10 28 Pri D

L1-4b 34 4 Se M

Women

C3b 23.04 7 Se S/P

C6b 21.11 6 Se S/P

T5a 44.10 1.10 Se M

T7a,c 53 6 Pri M

T9a 38.08 20 Se S

T10b,c 39.01 3 Pri D

T12a 41.11 19 Pri M

Pri, primary; Se, secondary; Su, superior; S, single; M, married; D, divorced; S/P, single with partner.a Extent of the injury: complete.b Extent of the injury: incomplete.c Incomplete sensory loss.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269258

3.4. Apparatus

The apparatus was that used in the pilot study. In addition, a Letica GSR-200

amplifier and a Letica CAR-300 amplifier were used to record skin conductance and

heart rate, respectively.

3.5. Physiological variables

Eyeblink Magnitude was recorded as in the pilot study. Heart rate was derived

from the blood pulse amplitude recorded at the distal phalanx of the right index

finger, using a Letica photoelectric sensor at a sensitivity of 1�/10 mV/full scale. The

inter-beat interval was measured in milliseconds and transformed into average heart

rate every half-second using a weighted averaging procedure (Reyes del Paso and

Vila, 1998). The heart rate response was expressed as change in heart rate from

baseline (before each picture) every half-second during the 6 s picture presentation.Skin resistance was measured from the second phalanx of the index and middle

fingers of the left hand, using standard Sensormedic electrodes filled with isotonic

electrolyte paste (0.29 g NaCl per 100 ml water). The amplifier used a constant

current of 10 A. Recording was done at a sensitivity of 100�/200 kV/full scale. Skin

resistance was transformed into skin conductance units and expressed as change in

conductance from stimulus onset to maximum increase during the 6 s slide

presentation.

3.6. Subjective measures

Valence and arousal ratings of the pictures were obtained using the SAM as in the

pilot study.

3.6.1. Structured interview

This interview included questions related to the following emotions: joy, love,

sentimentalism, fear, anger, and sadness. Participants were asked to search formemories of important and intense experiences in their past (always before the lesion

in the patient group) and present. Using the scale employed by Hohmann, they were

asked if they perceived each emotion less (1), the same (2), or more (3) than in the

past. For the analysis, the emotions of joy, love, and sentimentalism were grouped

into the category of ‘positive emotions’, whereas the emotions of fear, anger, and

sadness were grouped into the category of ‘negative emotions’.

3.7. Procedure

Recruitment of the spinal cord injury patients for the study was made by

contacting the Base Center for the Disabled of the province of Malaga (Spain) and

all the local Disabled Associations in the same province. The control participants

were selected from members of the staff of the faculty of psychology or from

contacts fulfilling the necessary requirements. In all cases, the laboratory session and

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 259

the interview were conducted by the first author, who waited for the participants at

the entrance to the faculty to take them to the laboratory. The atmosphere for the

session was always relaxed and unhurried, with no objections being raised by any

participant at any point, the patients collaborating fully. After completing the

informed consent form, the participant reclined in a comfortable chair and the

physiological sensors were attached. The instructions given to the participants were

the same as in the pilot study. Finally, after finishing the task and evaluating thepictures using the SAM, the participant completed the structured interview.

3.8. Data analysis

Analysis of variance was applied to the physiological variables using a factorial

design consisting of the group factor and the following repeated measures factors:

valence of the pictures (pleasant, neutral and unpleasant) and, when applicable,

noise (pictures with noise versus pictures with no-noise), and form of the response

(the physiological values within the 6 s picture presentation). As in the pilot study,

the Greenhouse Geisser epsilon correction was applied to the repeated measure

factors. Subjective variables were analyzed using the Friedman test (x2) for within-subject comparisons and the Student’s t-test or analysis of variance for between-

group comparisons. In addition, in order to directly compare our results with those

of Hohmann, an analysis was also undertaken of the frequency of patients reporting

decrement, no change, or increment in each emotional scale.

4. Results

4.1. Subjective ratings (SAM)

Ratings of valence and arousal showed significant differences between pleasant,

neutral, and unpleasant pictures for both SCI and control participants: valence in

SCI (x2(2, 19)�/38, PB/ 0.0001); valence in control ((x2(2, 19)�/38, PB/ 0.0001);arousal in SCI (x2(2, 19)�/26.6, PB/ 0.0001); arousal in control (x2(2, 19)�/22.92,

PB/ 0.0001). Average valence ratings were in line with the a priori selection of the

pictures: pleasant (SCI�/7.51, S.D.�/0.86; control�/7.83, S.D.�/0.6), neutral

(SCI�/4.45, S.D.�/0.75; control�/4.51, S.D.�/0.43), and unpleasant (SCI�/1.63,

S.D.�/0.59; control�/1.66, S.D.�/0.66). No significant differences were found

between the SCI and control participants. Average arousal ratings did not

significantly differ between the groups: pleasant (SCI�/4.85, S.D.�/2.01; con-

trol�/3.89, S.D.�/2.51), neutral (SCI�/3.67, S.D.�/1.74; control�/3.62, S.D.�/

1.52), and unpleasant (SCI�/7.37, S.D.�/1.31; control�/7.32, S.D.�/1.37). As in

the pilot study, the valence ratings showed significant differences between the three

valence categories whereas the arousal ratings showed significant differences only

between the unpleasant and the other two categories and not between the pleasant

and neutral ones.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269260

4.2. Eyeblink

Fig. 2 plots the eyeblink magnitude as a function of the groups and of the affective

valence of the pictures. Results of the 2�/(3) ANOVA (Group�/Valence) showed no

significant effects of the two main factors or of their interaction. As illustrated in Fig.

2, the expected valence effect was not observed. There was a non-significant tendency

for the SCI group to produce greater eyeblink magnitudes to all pictures compared

with the control group (F (1, 36)�/1.74, P �/0.10).

In order to test whether the absence of affective startle modulation in both groups

was due to the low arousal rating of the pleasant pictures, we reanalyzed the EMG

data and included only the three most arousing pleasant pictures, the three least

arousing neutral pictures, and the three most arousing unpleasant pictures. There

was a very slight tendency to show greater startle magnitude to the unpleasant

pictures than to the pleasant ones in both groups, but these differences did not

approach statistical significance (F (2, 36)�/0.95, P�/ 0.25).

4.3. Heart rate

4.3.1. Baseline level

Heart rate at baseline, before starting the pictures, was higher for SCI patients(80.99 bpm) than for control participants (75.24 bpm), although the difference did

not reach significance (F (1, 36)�/2.49, P�/ 0.10).

4.3.2. Response

Fig. 3 illustrates the heart rate response as a function of the groups and of the

affective valence of the pictures. Results of the 2�/(3�/2�/12) ANOVA (Group�/

Valence�/Noise�/Form) showed a significant main effect of the Form factor (F (11,

396)�/3.3, PB/ 0.05) and an interaction effect of the Valence�/Form factors (F (22,

792)�/3.06, PB/ 0.05). None of the other main factors and interactions were

Fig. 2. Eyeblink magnitude to the startle noise during pleasant, neutral and unpleasant picture viewing in

SCI and control groups. Error bars are one standard error.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 261

significant. As shown in Fig. 3, the general form of the heart rate response was

decelerative after the first 2 s. The differences in the form of the response as a

function of the valence of the pictures appeared in the last 3 s. Post hoc analyses

showed significant differences between the unpleasant and pleasant pictures from

second 3.5 to 6 (Tukey test: second 3.5, PB/ 0.05; second 4, PB/ 0.0001; second 4.5,

PB/ 0.0001; second 5, PB/ 0.0001; second 5.5, PB/ 0.01; second 6, PB/ 0.01), the

unpleasant pictures producing greater deceleration than the pleasant ones. The

neutral pictures had an intermediate decelerative form (between the pleasant and the

unpleasant ones), but post hoc analyses did not show any significant differences

between neutral and either pleasant or unpleasant pictures. The pattern of results did

not differ for the SCI and control groups. Although there was a tendency in the

patient group to show the modulatory pattern attenuated, none of the Group main

and interaction effects approached significance.

Fig. 3. Heart rate change during pleasant, neutral and unpleasant picture viewing in SCI and control

groups.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269262

4.4. Skin conductance

4.4.1. Baseline level

Skin conductance at baseline was lower for SCI patients (6.67 microsiemens) than

for control participants (9.07 microsiemens), although the difference did not reach

statistical significance (F (1, 36)�/2.90, PB/ 0.10).

4.4.2. Response magnitude

Results of the 2�/(3�/2) ANOVA (Group�/Valence�/Noise) showed significantmain effects of Group (F (1, 36)�/5.63, PB/ 0.05) and Noise (F (1, 36)�/23.89, PB/

0.0001) and the following significant interaction effects: Group�/Noise (F (1, 36)�/

4.45, PB/ 0.05), Noise�/Valence (F (2, 72)�/6.54, PB/ 0.01), and Group�/Noise�/

Valence (F (2, 72)�/4.81, PB/ 0.05). All significant effects are summarized in the

Group�/Noise�/Valence interaction represented in Fig. 4. Analysis of this

interaction showed a significant valence effect only in the control group, both

when the pictures were presented with noise (F (2, 36)�/5.46, PB/ 0.01) and when

they were presented without noise (F (2, 36)�/3.50, PB/ 0.05). Post hoc analysesshowed significant differences between neutral and unpleasant pictures in the noise

condition (Tukey test: PB/ 0.01), the unpleasant pictures producing the lowest

increase, and between the pleasant and unpleasant pictures in the no-noise condition

(Tukey test: PB/ 0.05), the unpleasant picture producing the greatest increase.

Fig. 4. Skin conductance change during pleasant, neutral and unpleasant picture viewing, under noise and

no-noise conditions, in SCI and control groups. Error bars are one standard error.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 263

4.5. Self-reported emotions

Most patients noticed either no change or an increase in their emotions, with just a

few feeling they had decreased. Some participants were unable to compare, and their

responses were considered as missing values in the analysis. When only the patient

group is considered (as in the Hohmann study), significant differences appear in the

scales of anger (decrease�/0; no change�/8; increase�/8; x2�/8.01, PB/ 0.05) andsadness (decrease�/1; no change�/5; increase�/11; x2�/8.95, PB/ 0.05), always in

the direction of increment. When compared with the control group, significant

differences appeared only in the sadness scale (t (1, 34)�/2.14, PB/ 0.05), with the

patient group showing a greater increment.

To determine the existence of differences in emotion assessments between the

subgroups of patients with high, middle or low lesions, the corresponding statistical

analyses were carried out. There was a general tendency in the group with middle

lesions to score higher for both pleasant and unpleasant emotions, although in nocase did these attain statistical significance.

The scores of patients with spinal cord lesions were compared according to the

extent of the lesion. There was a tendency for the group with incomplete lesions to

have higher scores in joy, love, and sentimentalism, and lower scores in anger, fear,

and sadness, although these were not statistically significant.

Finally, patient scores were compared according to the duration of the lesion.

There was a tendency for the group with longer duration to have lower scores in joy,

love, and fear, and higher scores in sentimentalism, anger, and sadness, althoughstatistical significance was not reached.

5. Discussion

In the present study, the subjective valence and arousal ratings of the pictures were

similar among SCI patients and control participants. They did not differ from the

normative ratings based on Spanish student samples (Molto et al., 1999). Thus, noevidence of reduced subjective experience to affective visual pictures was found in the

patient group. In addition, the patients and controls showed a similar affective

modulation of heart rate: both groups showed a significantly greater heart rate

deceleration while viewing unpleasant pictures than while viewing pleasant pictures,

with the neutral pictures producing intermediate heart rate responses. No evidence of

startle modulation was found in either group. However, evidence of skin

conductance modulation was found in control participants in interaction with the

noise condition: when the pictures were presented without noise, the unpleasant(most arousing) pictures produced greater skin conductance responses than the

pleasant and neutral ones, and the opposite was observed for picture�/noise

presentations. In general, these results are in agreement with those found by

Birbaumer (2001) in completely paralyzed ALS-patients using the picture viewing

together with the thought-translation-device for communication with the patients.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269264

Both physiological (heart rate and eyeblink startle) and subjective measures showed

no significant differences between patients and well controls.

The group with spinal cord injury scored higher than the control group in self-

reports of negative emotions, significantly so in report of sadness. Comparison

within the spinal cord injury group, according to the level, extension, and duration of

the lesion, provided no evidence for a decreased affective experience in patients with

greater peripheral loss. In general, these results are in agreement with those found by

Lowe and Carroll (1985), Chwalisz et al. (1988) and Bermond et al. (1991) in SCI

patients using Hohmann’s methodology. These three studies reported either no

change or an increase in emotional experience after the lesion, failing to replicate

Hohmann’s data. In general, these results clearly contradict James’ theory (1884)

and the so-called strong and weak forms of the feedback arousal theory (Chwalisz et

al., 1988; Reisenzein, 1983). According to the latter theory, peripheral physiological

feedback is either absolutely necessary for the emotional experience (strong form) or

is a contributing factor enhancing the intensity of this experience (weak form).

Our physiological data, however, showed differential effects that require

individual consideration. Our heart rate data replicate prior findings in well

participants: a greater final deceleration during the 6-s unpleasant pictures compared

with pleasant and neutral pictures (Bradley et al., 1993; Lang et al., 1993; Hare, 1973;

Klorman et al., 1977). This greater decelerative pattern to unpleasant visual stimuli

was interpreted by Lang and colleagues as indicative of a vagally mediated aversive

bradycardia similar to that found in passive aversive conditioning (LeDoux, 1990;

Obrist, 1981). The similarity between our two groups in the pattern of heart rate

affective modulation supports this interpretation, given that the vagal control of the

heart is not affected by the spinal lesion, whereas the sympathetic control (depending

on the level and extent of the lesion) is. Although there were nonsignificant group

effects in heart rate, the tendency to show an attenuated modulatory pattern in the

patient group might reflect this reduction in autonomic nervous system control on

the heart via the sympathetic branch. An alternative explanation would point to the

reduced number of participants in the study. A larger sample of both patients and

controls might have increased statistical power and the observed tendency would

have reached statistical significance.

The absence of affective eyeblink modulation in the main study, may have

multiple causes, including, among others, the age or type of the participants, or

inadequate startle trials or picture selection. The age of our participants in the main

study was considerably higher than in the pilot study, where affective modulation

was detected using the same task. The heterogeneity of our patients and control

participants regarding age and other characteristics might have contributed to

obscure the expected effect. Moreover, our task used a reduced number of trials (five

per affective category) and the pictures were selected exclusively on the basis of

valence ratings. Post hoc analyses showed that our selected pleasant pictures were

rated significantly lower in arousal than the unpleasant ones and that there were no

differences in arousal between the pleasant and the neutral pictures. Because the

modulation augments with increases in arousal for both pleasant and unpleasant

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 265

pictures (Cuthbert et al., 1996), the low arousal of the pleasant pictures might have

contributed to the absence of eyeblink modulation in our main study.

The skin conductance response to the pictures showed a significant reduction in

the patient group, with no evidence of affective modulation, which was only

observed in the control participants and in interaction with the noise condition. The

expected skin conductance modulation is a greater response to the most arousing

pictures, both pleasant and unpleasant (Lang, 1995; Bradley, 2000). The greaterresponse, under no-noise presentation, to the unpleasant pictures in our study,

compared to the neutral and pleasant ones, can be explained by the arousal ratings

of our pictures, which were significantly lower for the neutral and pleasant pictures.

The opposite effect for picture�/noise presentation is difficult to explain on the basis

of published results. Nevertheless, the strongest effect was the significant reduction

in the skin conductance response of the patient group. Because skin conductance is

exclusively mediated by the sympathetic branch of the autonomic nervous system,

which is affected by spinal cord lesion, we can speculate that the differences inaffective modulation observed between heart rate and skin conductance are

exclusively due to vagal influences. If this is so, our skin conductance data indirectly

support the above-mentioned interpretation of the greater decelerative heart rate

response to unpleasant pictures as a vagally mediated aversive bradycardia.

A theoretical problem in the interpretation of the physiological data is to what

extent the neural lesion in the patient group prevents any demonstration (against or

in favor) of central emotional processing using peripheral physiological measures

that are potentially affected by the lesion. Is the lack of affective modulation in thephysiological response an index of reduced emotional processing or just an index of

the peripheral lesion itself? In our study, this problem is especially relevant in the

skin conductance data, because the patient group showed a significant reduction in

this physiological response. Given the heart rate results, we are inclined to interpret

the skin conductance data as a direct consequence of the lesion rather than as an

absence of modulation. A related issue is to what extent different physiological

responses can show different patterns of modulation to the same affective stimuli.

Can eyeblink, heart rate, and skin conductance show different patterns ofmodulation? It should be noted that the physiological mechanisms of somatic

reflexes are quite different from those of autonomic responses. Moreover, within

autonomic responses, the mechanisms of heart rate are quite different from those of

skin conductance. Given these considerations, different patterns of modulation to

the same perceptual stimuli can be expected when different physiological indices are

used.

Finally, we would like to comment on some potential implications of our results

for traditional theories of emotion, in particular for the peripheralism versuscentralism controversy. On one hand, our self-report data suggest that emotions are

the results of central processes activated in the brain and that the subjective

component is probably independent of peripheral physiology. On the other hand,

some peripheral physiology, in particular heart rate, can index central emotional

processing, even in people with impaired somatic and autonomic physiology. Given

the modest correlations normally found between physiological and subjective

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269266

measures of emotions (Lang, 1995), peripheral physiological evidence that corrobo-

rates the normal affective reports of patients is not only a further argument against

the idea of impaired emotional processing in participants with impaired peripheral

physiology, but also an argument against the idea of emotion as a cognitive process

totally unrelated to bodily responses. The heart rate data in our study, like the

eyeblink response in Birbaumer’s study, clearly show a specific physiological

response pattern that covariates with self-reported emotion, even in people withsevere peripheral restriction. As Lang’s theory implies, an integrative approach to

emotion, which considers that emotion involves processes that are simultaneously

central and peripheral, seems more appropriate than the traditional dichotomous

approach of peripheralism versus centralism.

Acknowledgements

Support for this research was provided by the Spanish Ministry of Education

(research project PB97-0841) and Junta de Andalucıa (research group HUM 388).

References

Bermond, B., Nieuwenhuyse, B., Fasotti, L., Schuerman, J., 1991. Spinal cord lesions, peripheral

feedback, and intensities of emotional feelings. Cognition and Emotion 5, 201�/220.

Birbaumer, N, 2001. Emotion, brain activation and complete paralysis: fMRI and EEG as communication

devices. Abstracts of the II International Workshop on Emotion and the Brain. Universidad de las

Islas Baleares.

Birbaumer, N., Ohman, A. (Eds.), The Structure of Emotion: Psychophysiological, Cognitive, and Clinical

Aspects. Hogrefe & Huber, Seatle 1993.

Birbaumer, N., Ghanayim, N., Hinterberger, T., Iversen, I., Kotchoubey, B., Kubler, A., Perlmouter, J.,

Taub, E., Flor, H., 1999. A spelling device for the paralysed. Nature 398, 297�/298.

Bradley, M.M., 2000. Emotion and motivation. In: Cacciopo, J.T., Tassinary, L.G., Berntson, G.G.

(Eds.), Handbook of Psychophysiology. Cambridge University, New York, pp. 602�/642.

Bradley, M.M., Lang, P.J., Cuthbert, B.N., 1993. Emotion, novelty, and the startle reflex: habituation in

humans. Behavioral Neuroscience 107, 970�/980.

Cannon, W.B., 1914. The interrelations of emotions as suggested by recent psychological researches.

American Journal of Psychology 25, 256�/282.

Cannon, W.B., 1927. The James�/Lange theory of emotion: a critical examination and an alternative

theory. American Journal of Psychology 39, 106�/124.

Cannon, W.B., 1929. Bodily Changes in Pain, Hunger, Fear and Rage. Appleton, New York.

Cuthbert, B.N., Bradley, M.M., Lang, P.J., 1996. Probing picture perception: activation and emotion.

Psychophysiology 33, 103�/111.

Chwalisz, K., Diener, E., Gallagher, D., 1988. Autonomic arousal feedback and emotional experience:

evidence from the spinal cord injured. Journal of Personality and Social Psychology 54, 820�/828.

Cobos, P. (1999): Estados emocionales y patrones psicofisiologicos en la discapacidad: Lesion Medular.

Malaga (Spain): Servicio de Publicaciones de la Universidad de Malaga.

Dana, C.L., 1921. The anatomic seat of the emotions: a discussion of the James�/Lange theory. Archives

of Neurology and Psychiatry 6, 634�/639.

Fehr, F.S., Stern, J.A., 1970. Peripheral physiological variables and emotion: The James�/Lange theory

revisited. Psychological Bulletin 74, 411�/424.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 267

Garcıa, C., Jimenez, F., Cobos, M.P., Valero, L., 1997. BioLab: un programa informatico para la

adquicision de senales fisiologicas. Psicologica 18, 139�/151.

Graham, F.K., 1979. Distinguishing among orienting, defense and startle reflexes. In: Kimmel, H.D., Van

Olst, E.H., Orlebeke, J.F. (Eds.), The Orienting Reflex in Humans. Erlbaum, New Jersey.

Hare, R.D., 1973. Orienting and defensive responses to visual stimuli. Psychophysiology 10, 453�/463.

Hohmann, G.W., 1966. Some effects of spinal cord lesion on experienced emotional feelings.

Psychophysiology 3, 526�/534.

James, W., 1884. What is an emotion? Mind 9, 188�/205.

James, W., 1890. The Principles of Psychology. Holt, New York.

James, W., 1894. The physical basis of emotion. Psychological Review 1, 516�/529.

Klorman, R., Weissbert, R.P., Wiessenfeld, A.R., 1977. Individual differences in fear and autonomic

reactions to affective stimulation. Psychophysiology 14, 45�/51.

Kubler, A., Kotchoubey, B., Kaiser, J., Wolpaw, J., Birbaumer, N., 2001. Brain�/computer communica-

tion: unlocking the locked in. Psychological Bulletin 127, 358�/375.

Landis, C., Hunt, W.A., 1939. The Startle Pattern. Farrar, New York.

Lang, P.J., 1979. A bio-informational theory of emotional imagery. Psychophysiology 16, 495�/512.

Lang, P.J., 1985. The cognitive psychophysiology of emotion: fear and anxiety. In: Tuma, A.H., Maser, J.

(Eds.), Anxiety and the Anxiety Disorders. Erlbaum, Hillsdale, NJ.

Lang, P.J., 1994. The motivational organization of emotion: Affect-reflex connections. In: VanGoozen, S.,

Van de Poll, N.E., Sergeant, J.A. (Eds.), Emotions: Essays on Current Issues in the Field of Emotion

Theory. Erlbaum, Hillsdale, NJ.

Lang, P.J., 1995. The emotion probe: studies of motivation and attention. American Psychologist 50, 372�/

385.

Lang, P.J., Bradley, M.M., Cuthbert, B.N., 1997. Motivated attention: affect, activation and action. In:

Lang, P.J., Simons, F.R., Balaban, M.T. (Eds.), Attention and Orienting: Sensory and Motivational

Processes. Erlbaum, Hillsdale, NJ.

Lang, P.J., Davis, M., .Ohman, A., 2000. Fear and anxiety: animal models and human cognitive

psychophysiology. Journal of Affective Disorders 61, 137�/159.

Lang, P.J., Greenwald, M.K., Bradley, M.M., Hamm, A.O., 1993. Looking at pictures: affective, facial,

visceral, and behavioral reactions. Psychophysiology 30, 261�/273.

Lang, P.J., .Ohman, A., Vaitl, D., 1988. The international affective picture system (Photographic slides).

University of Florida, Center for Research in Psychophysiology, Gainesville, FL.

LeDoux, J.E., 1990. Information flow from sensation to emotion plasticity in the neural computation of

stimulus value. In: Gabriel, M., Moore, J. (Eds.), Learning and Computational Neuroscience,

Foundations of Adaptive Network. Bradford Books, Cambridge.

Lowe, J., Carroll, D., 1985. The effects of spinal injury on the intensity of emotional experience. British

Journal of Clinical Psychology 24, 135�/136.

Molto, J., Montanes, S., Poy, R., Segarra, P., Pastor, M.C., Tormo, M.P., Ramırez, I., Hernandez, M.A.,

Sanchez, M., Fernandez, M.C., Vila, J., 1999. Un nuevo metodo para el estudio experimental de las

emociones: El International Affective Picture System (IAPS). Adaptacion espanola. Revista de

Psicologıa General y Aplicada 52, 55�/87.

Obrist, P.A., 1981. Cardiovascular Psychophysiology: A perspective. Plenum, New York.

Pavlov, I.P., 1927. Conditioned Reflexes. Dover Publications, New York.

Reisenzein, R., 1983. The Schachter theory of emotion: two decades later. Psychological Bulletin 94, 239�/

264.

Reyes del Paso, G., Vila, J., 1998. The continuing problem of incorrect heart rate estimation in

psychophysiological studies: An off-line solution for cardiotachometer users. Biological Psychology 48,

269�/279.

Richards, J.S., Hirt, M.J.S., Melamed, L., 1982. Spinal cord injury: a sensory restriction perspective.

Archives of Physical Medicine and Rehabilitation 63, 195�/200.

Trieschmann, R.B., 1980. Spinal Cord Injuries: Psychological, Social and Vocational Adjustment.

Pergamon Press, New York.

Tucker, S.J., 1980. The psychology of spinal cord injury: Patient-staff interaction. Rehabilitation

Literature 41, 114�/121.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269268

Vila, J., Sanchez, M., Ramırez, I., Fernandez, M.C., Cobos, P., Rodrıguez, S., Munoz, M.A., Tormo,

M.P., Herrero, M., Segarra, P., Pastor, M.C., Montanes, S., Poy, R., Molto, J., 2001. El Sistema

Internacional de Imagenes Afectivas (IAPS): Adaptacion Espanola. Segunda parte. Revista de

Psicologıa General y Aplicada 54, 635�/657.

Wolpaw, J.R., McFarland, D.J., Vaughan, T.M., 2000. Brain computer research at the Wadsworth

Center. IEEE Transactions on Rehabilitation Engineering 8, 222�/226.

P. Cobos et al. / Biological Psychology 61 (2002) 251�/269 269