© 2002 American Thoracic Society
P3-Specific Amplitude Reductions to Respiratory and Auditory Stimuli in Subjects with AsthmaMusculoskeletal Research Centre, La Trobe University; Department of Psychology, The University of Melbourne, Melbourne, Australia; and Human Sleep Research Program, SRI International, Menlo Park, California Correspondence and requests for reprints should be addressed to Ian M. Colrain, Ph.D., SRI International, 333 Ravenswood Avenue, Menlo Park, CA, 94025. E-mail: ian.colrain{at}sri.com
The respiratory-related evoked potential (RREP) is the electroencephalographic response to brief inspiratory occlusion. The P3 component of the response reflects the active cognitive processing of stimulus information. This study investigated the RREP in 16 subjects with asthma and 16 control subjects. In addition, auditory-evoked potentials were recorded using an oddball paradigm in both groups and were compared with the RREP results. Perceptual sensitivity was assessed using a load magnitude estimation task, whereby subjects estimated the magnitude of four added resistive loads using a modified Borg scale. For RREP assessment, inspiratory occlusions were presented for 200 milliseconds. For auditory-evoked potential assessment, 1,000- and 2,000-Hz pure tones were presented at 60 dB for 100 milliseconds, with the 2,000-Hz tone presented with 20% probability and identified as a target stimulus. Scalp electroencephalographic activity was measured from 29 channels. Resistive load intensity was linearly related to magnitude estimation in both subjects with asthma (r2 = 0.96) and control (r2 = 0.99) subjects. Results showed that both the respiratory and auditory P3 components were markedly reduced in the group with asthma compared with the control group. Other components were similar between the groups. These results suggest an intrinsic reduction in P3 amplitude for patients with asthma, which may relate to differences in processing perceptual information.
Key Words: P3 • respiratory-related evoked potential • auditory-evoked potential • magnitude estimation • asthma
Failure to recognize the severity of an asthma attack is considered to be a common risk factor associated with asthma fatality (1–4). However, the reasons that some patients fail to recognize the severity of their asthma are not straightforward, as both psychological and physiological factors may be involved. Many psychophysical studies of respiratory load perception have been conducted in subjects with asthma using bronchial provocation or the application of external resistance to airflow. These studies show considerable variation between subjects for detection and magnitude estimation of both internal and external resistive loads and the perceived effort of breathing (for review see Ref. 5). In general, subjects with asthma require a greater added resistance to breathing than subjects without asthma for detection to occur. However, this difference disappears when data are normalized for the state of the background resistance before testing (6). In subjects with asthma, individual differences in the perception of symptoms do not correlate with obvious clinical markers, such as severity or duration of asthma symptoms (7–9). However, recent reports demonstrate that there is a subpopulation of asthma patients with an impaired ability to sense respiratory loads (10, 11), and the inclusion of these patients in the average may explain the lack of clinical correlation. In addition, this subpopulation of subjects with asthma with a reduced perceptual sensitivity to inspiratory loads has been suggested to have an increased likelihood of experiencing life-threatening asthma (10, 11). This is an important finding as it points to perception as being a key factor in the overall management of asthma. Regardless of the sensory input involved in the perception of asthma, a major factor explaining the wide differences in perception seen between individuals must be that of central processing. For example, the impairment in perception of bronchoconstriction in older people may be explained by impairment in central processing with age (1, 12). Furthermore, the sensation of dyspnea seems to depend on central rather than peripheral receptors (13). The activation of cortical neurons by mechanical loads has been studied using evoked potential techniques. In a pioneering study, Davenport and colleagues elicited a series of respiratory-related cortical evoked potentials in human subjects following inspiratory occlusion (14). This response was labeled the respiratory-related evoked potential (RREP). The amplitude of the first peak (P1) is maximal at the scalp region over the somatosensory cortex (15–19). It shows good correlation with psychophysical measures of respiratory sensation (17, 20, 21), and electromagnetic source estimation modeling has best described P1 as being produced by dipole generators in the primary somatosensory cortex of each hemisphere (18). Only one published RREP study has been performed in patients with asthma (22). This was performed in children with asthma, and results demonstrated an absent P1 component of the RREP in some of the children who were initially identified as being poor symptom perceivers because of a history of life-threatening asthma. It was suggested that such findings might be indicative of a dysfunction in the neural processing of respiratory load information in these patients. Importantly, this may explain why they have difficulty recognizing the presence or judging the severity of an asthma attack. It is therefore of interest to explore any deviations from normal RREP responses that may be a consequence of poor perception of respiratory load because the identification of neural markers of respiratory load sensation may serve to identify those patients most at risk of death from acute asthmatic episodes. In subjects without asthma, reduced sensitivity to respiratory effort has been associated with absent or atypical late RREP components (17), specifically the P3 component (a positive deflection with a latency in the range 300–600 ms). This is an interesting finding, as P3 does not appear to reflect the physical parameters of the eliciting stimuli as the earlier components do (e.g., P1) but rather, appears to reflect the active cognitive processing of stimulus information on the part of the subject. Although several studies have reported the presence of a P3 component in the RREP (15–17, 23, 24), no study has investigated the P3 or other long latency RREP components in patients with asthma. The primary aim of this study was to explore further the RREP in asthma. RREP components were recorded in both subjects with asthma and control adult subjects using a behavioral paradigm to allow for measurement of the P3 response. An estimate of each subject's perceptual ability was also made and compared with the RREP response. It was hypothesized that perceptual ability would have a positive correlation with RREP P3 magnitude. Auditory-evoked potentials (AEPs) were recorded in both groups and were compared with the RREP results. The purpose of this was to determine whether any evoked potential differences seen between subjects with asthma and control subjects were specific to respiratory stimuli or carried over to other stimulus modalities. It was hypothesized that greater differences between RREP and AEP components would be seen in the patients with asthma compared with the control subjects.
Subjects Sixteen control subjects were matched to 16 patients with asthma for age and sex. Each group consisted of eight males and eight females. The mean age ± SD of the asthma group was 22.1 ± 5.2 years, and the mean age of the control group was 21.2 ± 4.0 years. Mean body mass index was comparable between groups (asthma group body mass index = 23.6; control group body mass index = 21.9). Control subjects were nonsmokers with no known pulmonary disease. The patients with asthma met the American Thoracic Society's diagnostic criteria (25). Each patient had asthma since childhood and received daily maintenance treatment with inhaled corticosteroids to control their symptoms. The University of Melbourne's Human Subjects Ethics Committee approved the laboratory procedures. Subjects refrained from consuming alcohol 24 hours before their attendance. No medication restrictions were placed on the subjects with asthma.
Part 1: Magnitude Estimation of Resistive Loads A resistive load manifold was connected to the inspiratory line. This consisted of a rigid tube with a series of four side ports, between which sheets of stainless steel mesh (#300) were inserted to achieve inspiratory resistances of 0 (control), 3, 6, 12, and 24 cm H2O/L/second. A second computer determined the order of load presentation and digitally set the manifold via solenoid stop valves attached to each port. The manifold was set simultaneously with resistive load presentation, which was by manual activation of a trigger immediately before inspiration. The load was manually deactivated at the completion of one inspiratory effort by the same trigger, defaulting to zero resistance between load presentations. The manifold was positioned outside of the experimental room. Subjects were seated, and the respiratory apparatus was attached. The output of the airflow signal was connected to a purpose built device that consisted of a series of light-emitting diodes. Subjects were asked to close their eyes and breathe at their normal depth and rate while the experimenter adjusted the gain and zero-set controls of the light system such that the top light in the array of diodes was activated by the peak flow of each breath. This light system was placed in front of the subject who was instructed to ensure that the top light was activated with each breath. Subjects were provided with a verbal cue during the expiration preceding the breath for which they had to provide a load estimate. The subject made a written estimate immediately after the breath was complete using a modified Borg scale (26). Each load was presented 10 times in random order. Peak Pm and inspiratory flow were calculated from the recorded airflow and Pm signals and were compared between groups using an analysis of variance. The magnitude estimation results were averaged and plotted against resistive-load magnitude on a log–log scale. The linear regression equation was determined and tested for significance using analysis of variance.
Part 2: RREP Protocol Subjects breathed through the same respiratory apparatus as for Part 1, and a pressure-activated occlusion valve (Hans Rudolf series 2100) was connected to the inspiratory port of the nonrebreathing valve. The inspiratory occlusion was presented by manually activating the valve causing an interruption of inspiration within 250 milliseconds after the onset of inspiratory effort (occlusion methodology reported in detail previously [16]). Two to five unoccluded breaths separated each interruption occlusion. The inspiratory occlusion was approximately 200 milliseconds in duration. Subjects were required to maintain a constant flow rate throughout the recording period and to keep a count of each interruption they detected. In total, two hundred occlusions were presented. The 200 occlusions were averaged for each subject. An individual occlusion trial was included in the average if it produced a minimum 3 cm H2O change in Pm within 100 milliseconds of presentation. Five RREP components were defined as follows: P1 and Nf as the maximal positive and negative deflections 40–60 milliseconds following the commencement of the change in Pm, N1 as the next negative deflection after these early components 90–130 milliseconds, P2 and P3 as the following positive deflections in the ranges of 180–230 and 250–350 milliseconds, respectively (16, 17). Baseline to peak amplitudes were determined separately for each component at each electrode site. Latency values were expressed relative to the start of the Pm change and were obtained from the scalp site at which each component was maximal. Pm was measured from the initial maximum peak of the averaged waveforms. Comparisons between groups were performed with analysis of variance.
Part 3: AEP Protocol The 1000-Hz tone was presented with an 80% probability (standard) and the 2,000-Hz tone with a 20% probability (rare). The rare tone was designated as a target that the subject was instructed to count and make a button press in response to immediately upon detection. Reaction time was determined by the tone-generation software that recorded the difference in milliseconds between tone presentation and button press. Subjects did not respond to the standard tones. In total, two sets of 200-tone presentations were made. Subjects took a short break between sets. The standard and target tones were averaged separately. Four AEP components were defined as follows: N1 the most negative peak 90–130 milliseconds, P2 the most positive peak 150–250 milliseconds, N2 the most negative peak 200–300 milliseconds, and P3 the most positive peak 250–400 milliseconds poststimulus presentation (27). Baseline to peak amplitudes were determined separately for each component at each electrode site. Latencies were obtained from the scalp site at which the component was maximal. Analysis of variance was used to determine differences between stimulus type and group. For all parts, statistical significance was determined at the p < 0.05 level.
Part 1: Load Magnitude Estimation Pulmonary function data for the subjects with asthma are shown in Table 1. The mean age, height, and body weight was comparable between the subjects with asthma and the control subjects. The peak inspiratory airflow did not differ significantly between loads, F(4, 120) = 1.89, p > 0.05, and was not significantly different between subjects with asthma and control subjects, F(1, 30) = 0.48, p > 0.05. Thus, the subjects maintained peak inspiratory airflow despite the increasing resistance. The maximum Pm increased with increasing resistive load magnitude, F(4, 120) = 165.8, p < 0.001, and the relationship between Pm and inspiratory resistance was not significantly different between the subjects with asthma and the control subjects, F(1, 30) = 0.28, p > 0.05.
The mean Borg scores, shown in Figure 1 , were not significantly different between the groups, and progressive increases in the magnitude of resistive load were linearly related on a log–log scale to the estimated magnitude of the load in both groups. The slope of the line was 0.96 [R2 = 0.99, F(1, 2) = 159.1, p < 0.01] for the control group and was not significantly different from the slope of 0.87 [R2 = 0.96, F(1, 2) = 61.0, p < 0.05] for the asthma group. There was a wider range of slope scores for the group with asthma (0.32–1.50) compared with the control group (0.67–1.40).
Part 2: RREP Mouth occlusion pressure. There was no difference in initial peak Pm between the subjects with asthma (4.6 ± 1.2 cm H2O) and the control group (4.3 ± 0.5 cm H2O), t(30) = 0.45, p > 0.05.
RREP components.
There was no significant difference in latency for any component between the group with asthma and the control group (Table 2). Mean P3 amplitude was significantly larger [mean between group difference of 6 µV, F(1, 30) = 5.49, p < 0.05] in the control group compared with the asthma group. This difference can be clearly seen in Figure 2 , which shows grand mean evoked potential waveforms at the Pz electrode site, separately for the group with asthma and the control group. There was more variation in the amplitude of P3 for the subjects with asthma (range from 3 to 25 µV at Pz) as compared with the control subjects (range from 18 to 29 µV at Pz). The smallest P3 amplitude for the control subjects (18 µV) was larger than 11 of the 16 subjects with asthma. No other component showed a significant amplitude difference between the two groups.
Part 3: AEP N1, P2, N2, and P3 components were observed in all subjects. Grand mean amplitude and latency data are displayed in Table 3. Values for each component are displayed from the site that the component occurred maximally.
For both groups, P3 was reliably observed in only the target waveform. There was a significant difference in P3 amplitude in the target waveform between the groups, with the component being larger in the control group, F(1, 30) = 4.34, p < 0.05. This is highlighted by subtracting the standard waveform from the target waveform to produce a difference wave (see Figure 3) . There was considerable variation in the amplitude of P3 for both groups. The range of values for the subjects with asthma was 5 to 23 µV and 8 to 35 µV for the control subjects (measured at Pz).
The mean reaction time to the target tones was 525 milliseconds for the control group and 585 milliseconds for the group with asthma. This was, however, not statistically different, t(30) = 1.3, p > 0.05.
Part 4: Relationship between Magnitude Estimation and RREP P3 and P1 Amplitude
This is the first study to have investigated long latency RREP and AEP components in subjects with asthma. Five RREP components (Nf, P1, N1, P2, and P3) were reliably elicited in both groups following midinspiratory occlusions. Four AEP components were identified to the auditory tones in both groups (N1, P2, N2, and P3). The predominant finding was that the subjects with asthma had significantly reduced P3 amplitudes to both respiratory and auditory stimuli compared with the control subjects. The reasons that subjects with asthma as a group showed a markedly smaller P3 amplitude to both respiratory and auditory stimuli are not altogether clear. However, the smaller P3 amplitude was not due to an overall dampening of evoked potential amplitudes, as the N1 of the RREP and the P2 of the AEP are bigger in subjects with asthma, and all other components (other than P3) are the same. Thus, there appears to be a specific reduction of P3 amplitude for patients with asthma. These results also demonstrated a wide range of individual differences in the ability of the subjects to perform the load magnitude estimation task, consistent with previous reports (28). Although the slope of the magnitude estimation-resistive load relationship was not significantly different between the group with asthma and the control group, the subjects with asthma showed greater intersubject variability than the control group. Inspection of the relationship between the RREP P3 amplitude and the magnitude estimation slope (see Figure 4) reveals that subjects who had the smallest P3 amplitudes also had the lowest magnitude estimation slope. This is consistent with the findings in subjects without asthma in Webster and Colrain (17), in which within-subject linear relationships were found between P3 amplitude and magnitude estimates. These data indicate that the same relationship generally holds between subjects. The relationship between the ability to make subjective judgments about load and P3 cannot be used to explain the group differences in P3 amplitude, however. First, all of the subjects with asthma appeared to perceive increased resistive load reasonably well and were able to assign sensible scores using the Borg scale. Certainly, none demonstrated performance anything like that of the two poor perceiving subjects without asthma reported by Webster and Colrain (17). Second, as can be seen in Figure 4, at least three of the subjects with asthma with low P3 amplitude had slope values of greater than one. The fact that all subjects with asthma showed a normal pattern of early components suggests that the differences in P3 amplitude may be due to subsequent cortical processing of the signal rather than afferent transduction or transmission of the signal to primary cortical areas. This conclusion is, however, made with caution, as this is the first study to investigate the long latency evoked potentials in subjects with asthma and is also in contrast to that of Davenport and colleagues (discussed later here) (22). There has been only one reported study to investigate the RREP in asthma (22). Three groups of children were studied. Group 1 consisted of 11 children with a history of life-threatening attacks. Group 2 consisted of 15 children with asthma but no history of life-threatening events. Group 3 consisted of 14 children with no history of asthma. RREPs were recorded bilaterally from C3, C4 scalp locations, referenced to Cz, and elicited by midinspiratory occlusion. Results showed the P1 peak to be absent in 6 of the 11 patients in the life-threatening asthma group and was present on all but one of the remaining subjects (a patient with asthma that was being considered for inclusion in the high-risk program). The authors concluded that the absence of the RREP in the half of the children with a history of severe life-threatening asthma attacks suggested that the normal load-related sensory information was not reaching the somatosensory cortex. Given these findings of reduced P3 amplitude in subjects with asthma without life-threatening events, it would be interesting for future research to explore long latency RREP components in patients with life-threatening asthma. The results of this study are generally similar to a previous study that compared the detection of resistive loads, and RREPs to resistive loads between subjects with obstructive sleep apnea and age-matched controls (29). They reported no difference in the threshold resistance detected between the obstructive sleep apnea and control subjects. For the RREP study, three loads (3.6 cm H2O/L/second, 9.4 cm H2O/L/second, or total occlusion) were used. Five RREP components were identified (P1, N1, P2, N2, and P3) in both groups of subjects; however, the amplitude of P3 was significantly reduced in the obstructive sleep apnea patients compared with the control subjects when the load was in the form of a total occlusion. Similar to these results, the cause of this difference is unknown but may include such mechanisms as specific upper airway disturbances or the long-term effect of prolonged hypoxia on central processes. The finding that subjects with asthma have a reduced auditory P3 in addition to a reduced respiratory P3 is intriguing. One explanation for this finding is that both components are manifested via the same processes. In other words, if a deficit in cortical processing in one stimulus modality affects one (or more) of the P3-producing generators and therefore dampens the P3 response, the same effect will to some extent also be seen when recordings are made using a different stimulus modality. However, it is also possible that respiratory stimuli and auditory stimuli use different pathways and neural mechanisms. Thus, although a change in respiratory and not auditory responses would suggest a respiratory specific deficit, the alteration of both responses suggests a general sensory deficit. Therefore, in summary, the results of this experiment point to P3 as a potential index of the perceptual awareness of increased airflow resistance, which is generally consistent with the findings of Webster and Colrain (17). The RREP P3 may additionally provide insight to problems of symptom discrimination, given that these results indicated that patients with asthma with poor discrimination have dampened P3 responses. Thus, there does appear to be a difference in the mechanism of perceptual processing of respiratory information between subjects with and without asthma. The exact nature of this mechanism is, however, yet to be determined. The findings of this experiment will hopefully serve as a basis for subsequent reports of evoked potentials to respiratory obstruction in patient populations. Received in original form December 4, 2000; accepted in final form February 21, 2002
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
