Comparison of Thermogenic Sympathetic Response to Food Intake between Obese and Non-obese Young Women


Laboratory of Applied Physiology, Graduate School of Human and Environmental Studies, Kyoto University, Sakyo-ku, Kyoto 606, Japan. E-mail:


Objective: Sympathetic nervous system abnormality in humans is still a matter of debate. The present study was designed to examine diet-induced autonomic nervous system activity and metabolic change in obese and non-obese young women.

Research Methods and Procedures: Sixteen age- and height-matched obese and non-obese young women participated in this study. Sympathovagal activities were assessed by means of our newly developed spectral analysis procedure of heart-rate variability during the resting condition and after mixed-food ingestion (480 kcal). Energy expenditure was also measured under these two conditions.

Results: There was no significant difference in any of the parameters of the heart-rate variability between the obese group and control group during the resting condition. In the control group, both absolute values (221.5 ± 54.5 vs. 363.8 ± 43.7 ms2, p < 0.05) and relative values (0.23 ± 0.03 to 0.36 ± 0.02, p < 0.05) of a very-low-frequency component and global sympathetic nervous system index (1.46 ± 0.19 vs. 3.26 ± 0.61, p < 0.05) were significantly increased after mixed-food ingestion compared with the values obtained after resting condition. However, no such sympathetic response was found in the obese group. Energy expenditure increased in the two groups after the meal, but the magnitude of the increase above the preprandial resting condition was significantly greater in the control group than in the obese group (11.2 ± 2.3 vs. 6.7 ± 0.8%, p < 0.05).

Discussion: Our data suggest that despite identical sympathovagal activities at the resting condition, obese young women may possess a reduced sympathetic response to physiological perturbation such as mixed food intake, which might be related to lowered capacity of thermogenesis and the state of obesity.


Most individuals, whether lean or obese, maintain a constant body weight and body composition for prolonged periods of time. Despite wide fluctuations in physical activity and energy intake, the average drift in body weight has been estimated to be ∼1 kg/yr for males between the ages of 20 and 60 years (1) (2). A regulatory system that maintains constant energy storage is likely to involve complex interactions among humoral, neural, metabolic, and psychological factors, and it has been suggested that the autonomic nervous system (ANS) may be central in the coordination of this system (3).

As an important contributor to the regulation of energy expenditure (EE), the sympathoadrenal system is widely assumed to play a major role in the pathophysiology of obesity. The MONA LISA hypothesis, an acronym for Most Obesity kNown Are Low In Sympathetic Activity (4), has been strongly supported (5) (6) (7), whereas disagreement still exists over the nature of the sympathetic abnormality within the obese population (8) (9) (10) (11). This may be partly attributable to differences in methodology. The activity of the sympathetic nervous system (SNS) has been assessed in several ways, including plasma catecholamine concentration, catecholamine turnover, urinary catecholamine excretion, and muscle sympathetic nerve activity (12) (13). It has not been clarified, however, whether these measurements are suitable for evaluating the physiological function of the SNS, which especially contributes to the energy homeostasis.

Changes in parasympathetic nervous system (PNS) tone have also been implicated in the obesity of the Zucker rat (14) and the ventromedial hypothalamus-lesioned rat (15). Although a relationship has been found between PNS activity and total energy storage in humans (6), the specific involvement of the PNS in human energy metabolism is still unclear. Evaluation of PNS activity is experimentally more difficult to achieve because no direct measurement of PNS activity exists in humans at the present time. Animal studies have depended on invasive techniques such as counting the rate of firing of the vagus nerve, a technique that cannot be performed easily in humans.

Because both the SNS and PNS seem to be involved in modulating energy homeostasis, investigating the interactions between these two branches of the ANS may provide a new understanding of enigmatic human obesity mechanisms.

The beat-to-beat variation in heart rate (heart-rate variability [HRV]) has been a useful and safe tool to measure ANS activity and provides a comprehensive quantitative and qualitative evaluation of neuroautonomic function (16) (17) (18) (19). Power spectral analysis of HRV has shown at least two distinct regions of periodicity in electrocardiogram (ECG) R-wave intervals. In general, the high frequencies (HI) (>0.25 Hz) of HRV are associated solely with PNS activity, and the low frequencies (LO) (<0.10 Hz) of HRV might be associated with both SNS and PNS activities (16) (17) (20). Very low (VLO) frequency is not well described; however, it has been pointed out that frequencies much lower than 0.15 Hz may reflect the thermoregulatory vasomotor control system (21) (22) (23). We have demonstrated recently that VLO-frequency components (0.007 to 0.035 Hz), identified by our new mathematical technique, were selectively increased in humans against thermogenic perturbation such as acute cold exposure and spicy food containing capsaicin (24) (25) (26). This finding suggests the possibility of evaluating the SNS activities associated with energy metabolic regulation by means of HRV spectral analysis in humans.

Food intake has been shown to influence both SNS and PNS activities and to enhance thermogenic response in normal subjects (27) (28). Thus, investigating diet-induced ANS activity and metabolic change may be an appropriate approach to identify possible abnormalities in thermogenic autonomic response in the obese population. Accordingly, in the present study we evaluated the ANS activities in age- and height-matched obese and non-obese young women under the resting and postprandial conditions by means of our newly invented HRV spectral analysis (24) (25) (26) and investigated whether the sympathovagal as well as the metabolic responses to thermogenic perturbation were altered in obese young women.

Research Methods and Procedures


Sixteen age- and height-matched obese and non-obese young women were recruited for this study. After weighing each subject on a scale (BWB-627; Tanita, Tokyo, Japan), the percentage of body fat was determined by means of dual-energy X-ray absorptiometry (QDR = 1000; Hologic, Inc., Waltham, MA). Body mass index was also calculated as body weight divided by square height. Descriptive characteristics of the subjects are presented in Table 1.

Table 1.  Physical characteristics of subjects
 Control (n = 8)Obese (n = 8)
  • Values are means ± SE.

  • *

    p < 0.01.

Age (year)19.8 ± 0.8820.0 ± 0.33
Height (cm)158.8 ± 0.94159.9 ± 1.26
Weight (kg)47.0 ± 0.9774.5 ± 2.56*
Body mass index (kg/m2)18.6 ± 0.4029.0 ± 1.02*
% fat20.5 ± 1.2035.8 ± 1.37*
Fat mass (kg)9.6 ± 0.6526.8 ± 1.71*
Lean body mass (kg)37.3 ± 0.8447.7 ± 1.40*

The study protocol was approved by the Institutional Review Board of Kyoto University Graduate School. All subjects were carefully instructed about the study, and all gave their informed consent to participate in the study. Each subject completed a standardized health questionnaire for past medical history, medication, lifestyle, diet, smoking habits, alcohol consumption, and physical activity. All subjects were nonsmokers in good health and had no evidence of hypertension, cardiovascular disease, diabetes mellitus, or other endocrine diseases. All had been weight-stable for at least 1 year. Subjects were requested to avoid any medication for 1 week before the study and to keep to their usual diet. Each subject was instructed to avoid any food or beverage containing alcohol or caffeine after 9:00 pm of the day preceding the study.

Experimental Procedure

Subjects came to the laboratory at 9:00 am; all experiments were performed in the morning. The room was temperature-controlled (23 to 24°C) and quiet with minimization of arousal stimuli. ECG electrodes were placed on the subjects, who then rested for at least 20 minutes before the start of the experiment.

After the resting period, CM5 lead ECG and gas-exchange parameters, using an Aero-monitor AE 280 (Minato Medical Science, Tokyo, Japan), were continuously recorded while the subject remained seated in a comfortable chair for 20 minutes. The subject then ingested the mixed-food diet (480 kcal; 55% of the energy was in the form of carbohydrate, 15% was in the form of protein, and 30% was in the form of fat) over a 5-minute period. Immediately after this period, the ECG and gas-exchange measurements were performed continuously for 35 minutes. During resting and postprandial conditions, all subjects breathed in synchrony with a metronome at 15 beats/min (0.25 Hz) to ensure that respiratory-linked variations in heart rate did not overlap with low-frequency heart-rate fluctuations (<0.15 Hz) from other sources.

Identification of Sympathetic Component Related to Thermoregulation

In the present study, a newly invented HRV spectral analysis was used to evaluate and quantify sympathovagal activity. The process of developing the new technique of HRV spectral analysis has been fully described in our recent studies (24) (25) (26). In short, pharmacological blockade experiments (parasympathetic muscarinic blocker, atropine; β-sympathetic blocker, propranolol) were conducted on six healthy young subjects to examine the effects of sympathovagal activities on regulating energy metabolism in humans. When the ANS was blocked completely by these pharmacological agents, heart-rate fluctuations were almost entirely abolished and the resting EE was significantly reduced (−298.0 ± 47.8 kcal/day, p < 0.01) (24) (25).

To identify the spectral region associated with the thermogenic component of the SNS, the subjects were given physiological stimulation, such as cold exposure and food ingestion. The Simplex Method (29), which is a mathematical maneuver to decompose a nonlinear signal into its constituent elements, was used to detect the specific power spectral peaks associated with thermoregulation.

Figure 1 represents the lower-frequency components (<0.15 Hz) of power spectral data obtained from a healthy young subject during the resting condition at comfortable room temperature (25 °C, Figure 1A) and acute cold exposure (10 °C, Figure 1B). As Figure 1, A and B illustrate, by using the Simplex spectral peak detection method, the frequency region <0.15 Hz in the power spectral curve (thick line) was found statistically to consist of several peaks (thin lines). We observed that the lower spectral peak (<0.03 Hz) increased predominantly at the ambient temperature of 10 °C compared with 25 °C. Based on the data obtained from the subjects participating in the preliminary study, the integrated region of 0.007 to 0.035 Hz (an average spectral power peak of 0.021 ± 0.0047 Hz) was defined as the VLO-frequency component related to the thermoregulatory function of the SNS activity (Figure 1C) (24) (25) (26). It should be noted that the VLO-frequency component was significantly increased in the non-obese healthy young subject group after ingestion of capsaicin-containing spicy food as well as cold exposure (24) (25) (26). Therefore, the observations in our preliminary studies using pharmacological blockades and thermogenic perturbation suggest that the VLO-frequency component might serve as an indicator for the thermoregulatory sympathetic function.

Figure 1.

Examples of heart-rate variability power spectra for a healthy young volunteer (A) at the comfortable room temperature (25 °C) and (B) during acute cold exposure (10 °C). The simplex spectral peak detection method demonstrates the lower-frequency components (<0.15 Hz), statistically consisting of several peaks represented by thin lines. The lowest peak increased predominantly against acute cold exposure compared with the other peaks. (C) Power spectral components of HRV. Very-low- (0.007 to 0.03 Hz), low- (0.03 to 0.15 Hz), and high- (0.15 to 0.5 Hz) frequency components are represented by the black area, the lined area, and the white area, respectively.

R-R Spectral Analysis Procedure

Our R-R interval power spectral analysis procedures have been fully described elsewhere (18) (19) (24) (30). Briefly, analog output of the ECG monitor (Life Scope; Nihon Kohden, Tokyo, Japan) was digitized via a 13-bit analog-to-digital converter (HTB 410; Trans Era, South Orem, UT) at a sampling rate of 1000 Hz. The digitized ECG signal was differentiated, and the resultant ECG QRS spikes and the intervals of the impulses (R-R intervals) were stored sequentially on a hard disk for later analyses (18) (19).

Before R-R spectral analysis was performed, the stored R-R interval data were displayed and aligned sequentially to obtain equally spaced samples with an effective sampling frequency of 2 Hz (31) and displayed on a computer screen for visual inspection. Then, the direct current component and trend were completely eliminated by digital filtering for the band-pass between 0.007 and 0.5 Hz. The high-pass filtering at 0.007 Hz was chosen to include the frequency components associated with thermogenic functions of the ANS (24) (25) (26). The root mean square value of the R-R interval was calculated as representing the average amplitude. After passing through the Hamming-type data window, power spectral analysis by means of a fast Fourier transform was performed on consecutive 1024-second time series of R-R interval data obtained during the test. To evaluate ANS activity in each subject of the present study, we analyzed very low frequency (0.007 to 0.035 Hz), low frequency (0.035 to 0.15 Hz), high vagal component (0.15 to 0.5 Hz), and total power (0.007 to 0.5 Hz) by integrating the spectrum for the respective band width (Figure 1C). In addition, indices of the SNS and PNS activities were also calculated as the ratio of (VLO + LO)/HI and HI/TOTAL, respectively (18) (19) (24) (25) (26). We defined VLO and VLO/TOTAL as the absolute and relative thermogenic SNS activities, respectively. The mean heart rate of each 1024-second segment was also calculated with SD.

Calculation of EE

EE was determined from the oxygen (O2) consumption and respiratory quotient (RQ), calculated as the ratio of CO2 produced to O2 consumed using the following formula: EE (kcal/min) = (4.686 + [(RQ − 0.707) ÷ 0.293] × 0.361) × VO2, where 4.686 kcal/liter is the energy value of 1 liter of O2 at a nonprotein RQ of 0.707; RQ is the measured respiratory quotient; 0.707 is the RQ when only fat is oxidized; 0.293 is the difference between the RQ for carbohydrate and fat oxidation; 0.361 is the difference in energy value of a liter of O2 between an RQ of 1 and that of 0.707; and VO2 (liter/min) is the rate of oxygen consumption at standard temperature, pressure, dry conditions (32).

Statistical Analyses

All statistical analyses were performed using a commercial software package (SPSS version 7.5 for Windows; SPSS Inc., Chicago, IL). A Student's unpaired t test was performed to assess statistical differences in physical characteristics between the two groups. The effect of mixed-food ingestion was compared in the two groups by applying a two-way ANOVA with repeated measures to the parameters regarding ANS and EE. All p values <0.05 were considered to be statistically significant. Data are expressed as mean ± SE.


Figure 2 represents group data with respect to the R-R spectral parameters (VLO, VLO/Total, SNS Index, and PNS Index) obtained from eight control and eight obese subjects during the resting and postprandial conditions. It should be noted that during the resting conditions, there was no significant difference in any of the parameters of HRV including VLO, VLO/TOTAL, SNS Index, and PNS Index between the control and obese groups during the resting condition.

Figure 2.

Comparison of (A) very-low-frequency component, (B) the ratio of very-low-frequency component to total power (VLO/TOTAL), (C) SNS Index, and (D) PNS Index between resting and postprandial conditions, in the control and the obese groups, respectively. Results are expressed as mean ± SE for each group. *p < 0.05.

After eating the mixed-food diet, heart rate increased significantly both in the control group (71.5 ± 1.91 to 76.7 ± 2.19 beats per minute, p < 0.01) and the obese group (66.9 ± 2.56 to 70.6 ± 2.17 beats per minute, p < 0.05). The control group demonstrated a significant increase in VLO (221.5 ± 54.5 to 363.8 ± 43.7 ms2, p < 0.05), VLO/TOTAL (0.23 ± 0.03 to 0.36 ± 0.02, p < 0.05), and SNS Index (1.46 ± 0.19 vs. 3.26 ± 0.61, p < 0.05), and a significant decrease in PNS Index (0.43 ± 0.03 vs. 0.27 ± 0.04, p < 0.05) after the mixed-food diet. In contrast, the VLO, VLO/TOTAL, and SNS Index did not increase in the obese group in response to food ingestion. With regard to the PNS Index, no significant change was found in the obese group.

When the two groups were compared under the postprandial condition, the obese group demonstrated a significantly lower VLO/TOTAL (0.23 ± 0.05 vs. 0.36 ± 0.02, p < 0.05) and SNS Index (1.57 ± 0.44 vs. 3.26 ± 0.61, p < 0.05), and a higher PNS Index (0.47 ± 0.07 vs. 0.27 ± 0.04, p < 0.05) than the control group.

EE was increased in the control group (0.79 ± 0.02 to 0.90 ± 0.02 kcal/min, p < 0.01) and the obese group (1.00 ± 0.02 to 1.06 ± 0.03 kcal/min, p < 0.05) after the mixed-food ingestion. However, the magnitude of the increase above the preprandial resting condition was significantly greater in the control group compared with the obese group in response to the food ingestion (11.2 ± 2.3 vs. 6.7 ± 0.8%, p < 0.05).


The ANS, especially the sympathetic branch of the system, has been thought to contribute to the modulation of energy homeostasis and consequently to determination of the state of obesity (3) (4). Plasma or urinary norepinephrine estimates commonly have been used as global indices of sympathetic nervous activity; however, they have provided conflicting results over the nature of the SNS affected by obesity (an increase or a decrease) (6) (8) (9). More recently, several investigators have used microneurography to delineate muscle sympathetic nervous activity (MSNA), demonstrating a positive correlation between the amount of body fat and MSNA (10) (11). Because microneurography measures the efferent activity in sympathetic nerves supplying blood vessels in skeletal muscle, the study of MSNA in the control of blood pressure is quite logical; however, it may be less relevant to measure MSNA in studies in which the sympathetic control of energy metabolism rather than of the sympathetic control of blood pressure is of principal interest.

Concerning the thermogenic component of the ANS activity, it has been shown that catecholamine turnover within cardiac tissue correlates strongly with autonomic effects on energy metabolism elsewhere in the body (33). A previous study has shown that metabolic changes after glucose ingestion are associated with a predominant sympathetic activity in cardiac sympathovagal balance measured by HRV spectral analysis (34). We have recently invented a new technique of HRV spectral analysis that enables us to identify three frequency components (VLO, LO, and HI), and which demonstrates that the VLO-frequency component is selectively and remarkably increased after thermogenic perturbations among healthy individuals (24) (25) (26). These findings suggest that the HRV spectral analysis could be an appropriate method to explore the sympathothermogenic activity in humans.

Components of HRV have attracted considerable attention from investigators in various physiological fields, including diabetes and obesity research, and have become reliable noninvasive measures to evaluate sympathovagal activity (18) (19) (24) (25) (26) (34) (35) (36). Despite its prevalence and usefulness, quantification and interpretation of HRV remain complex issues and are fraught with pitfalls unless the neurophysiological concept of HRV is fully understood (37) (38). There is an array of variables (including age, subjects’ clinical features, and respiratory frequency) that should be controlled when adequately assessing the sympathovagal activity by way of HRV spectral analysis (37) (38). Concerning the experimental condition, the validity of spectral analysis of HRV has been proved at rest. Several studies, however, have faced a difficulty in using HRV spectral analysis during nonresting conditions such as dynamic exercise (39) (40). In this case, power in the frequency domain was reduced markedly because of the elevated heart rate with minimal HRV, leading to a problem of an extremely low signal-to-noise ratio, which makes spectral analysis unreliable.

In the present study, there was no significant difference in clinical features, except for the factors related to obesity, between the control and obese groups. During the experiment, the respiratory rate was carefully controlled at 15 beats/min (0.25 Hz) by an electric metronome to avoid the parasympathetic component interfering with the lower-frequency components. Furthermore, in contrast to dynamic exercise, heart rate was found to increase by ∼5 beats/min in both the control and obese groups in response to mixed-food ingestion. This rate of increase in heart rate would not affect analyses of HRV spectral components (19) (25) (30) (35).

Using our new device to investigate autonomic state, we have demonstrated that obese young women showed a lower VLO-frequency component than non-obese young women after food ingestion. As described above, the factors affecting the ANS activity were carefully controlled in the present study. In addition, the efficacy and reliability of this method have been demonstrated in our previous studies using pharmacological blockades and thermogenic stimulation (24) (25) (26). Taken together, our findings indicate that the VLO-frequency component might serve as an indicator of the SNS activity associated with thermoregulation and that the reduced VLO-frequency component found in the obese group implies a lowered diet-induced sympathetic response.

The lowered SNS response to food ingestion found in the obese young women is in agreement with Bazelmans et al. (41), who reported a decreased stimulation of norepinephrine appearance in obese subjects in response to overfeeding. A previous study of Paolisso et al. (34) using HRV spectral analysis has shown that the rise in the low- to high-frequency ratio after glucose ingestion negatively correlated with body fat content. Spraul et al. (42) have also found a negative correlation between the percentage of body fat and the increase in the MSNA in response to glucose ingestion, whereas fasting MSNA correlated positively with body fat. Recent findings in our laboratory (24) (25) (26) have shown that the VLO-frequency component of HRV as well as its responsiveness were markedly reduced in obese young women after the other physiological stimulations, such as cold exposure and ingestion of spicy food containing capsaicin. Although experimental designs are not always the same as that used in the present study, our data support these previous investigations and reinforce the MONA LISA hypothesis, indicating that obesity is associated with a relative or absolute reduction in the activity of the thermogenic component of the SNS.

In addition to the reduced SNS response to mixed-food ingestion, we have shown that the magnitude of increase in EE was significantly lower in the obese group compared with the control group. The ANS has been reported to influence thermogenic response to food intake in lean individuals. Astrup et al. (43) reported that β-adrenergic blockade reduces the carbohydrate-induced increase in forearm oxygen consumption. De Jonge and Garrel (28) have shown recently that sympathetic inhibition with propranolol decreased the “facultative” component of mixed-food-induced thermogenesis by 25%. Thus, reduced diet-induced thermogenenic capacity found in obese young women may be caused by the impaired activation of SNS related to thermoregulation.

Little information is available on the parasympathetic function in human obesity. In the present study, we found that the PNS activity was significantly decreased only in the control group during the early stage of the postprandial condition. According to Bray (4), the overall control of energy intake is integrated by reciprocal changes in the two divisions of the ANS. If that is the case, it could be speculated that the PNS activity might be suppressed temporarily by the rapid increase of the SNS activity in this stage. Such a parasympathetic reaction might not have appeared in the obese group due to the reduced response of the ANS. Yet because of the shorter duration of measurement, which was designed to investigate an acute ANS response to thermogenic perturbation, the possible physiological roles of the PNS activity in energy homeostasis could not be explained fully by the present study. Further research will be needed to elucidate the interaction between SNS and PNS activities in thermogenic responses to food intake and the contribution of PNS to the pathophysiology of human obesity.

In summary, we demonstrated that there was no significant difference in the resting ANS activity between obese and non-obese young women. After the mixed-food ingestion, however, the SNS activity associated with thermoregulation, as well as its response, was significantly lower in the obese group compared with the control group. Thus, our data suggest that the reduced sympathothermogenic response, which may cause diminished energy expenditure and further weight gain, could be an important etiologic factor leading to obesity in young women.


This work was supported by a Japanese Ministry of Education, Science, Sports and Culture Grant-in-Aid for Scientific Research (B) 11480011.