Urination Urge Effect on Heart Rate and Galvanic Skin Response
LiKang Chin, Richard Dela Rosa, Cristian Jurau, and Gregory Moy
Department of Bioengineering, University of Pennsylvania, Philadelphia, PA
Abstract
The urge to urinate stimulates the sympathetic nervous system, which triggers observable physiological responses. Change in heart rate (beats per minute) and galvanic skin response (Ohm-1) are two of these responses measurable by BioPacPro 3.6. They were analyzed to determine if there are significant correlations with this urge. It was found that the average heart rates, as well as their variances for seven out of nine subjects, increased significantly when the urge to urinate was present (p=0.05). The average heart rate across the entire subject population increased significantly (p=0.05) from 69.93 to 76.84 bpm. The variance of the heart rates across the total population increased with the urge to urinate as well (p=0.05). The average value and variance of the difference between the GSR values before and after urination for the entire subject population did not change significantly (p=0.05). The data shows that the variance for only two out of nine average individual GSR values increased significantly (p=0.05) under urination urge.
1. Background
The experimental goal is to investigate the effects of the urination urge on heart rate and galvanic skin response. Such a relation would be useful in clinical settings, where patients who are unable to communicate such an urge would be able to receive assistance via automated signaling.
The distension of the bladder due to accumulating urine is perceived as a urination urge and stimulates the sympathetic nervous system, which elicits generalized responses (Campbell 1012). Some of the common effects of sympathetic stimulation are contraction of the arteriolar smooth muscle and increase in perspiration. Contraction of the arteriolar smooth muscle causes an increase in heart rate, while perspiration increases skin surface water and electrolytes. The increase in electrolytes in turn lowers electrical resistance and thus increases the electrical conductivity of skin. (Pflanzer 3) Therefore, heart rate and skin conductivity were monitored by measuring EEG and GSR using the BioPacPro 3.6 package.
2. Materials
BIOPAC disposable vinyl electrodes (EL503)
BIOPAC Electrode lead set (SS2L)
BIOPAC Electrode gel (GEL1)
BIOPAC GSR transducer (SS3L)
Computer system (PC running Windows 95)
BIOPAC Pro Student Lab software V3.6
BIOPAC acquisition unit
BIOPAC wall transformer
1 Liter H2O
3. Methods
Channel 1 was used to monitor the ECG signal. AC input coupling along with a high pass hardware filter of 0.05 Hz, which is larger than the lowest expected heart rate, was chosen to increase resolution and screen out the low frequency noise. After analysis of a sample signal frequency content, two low pass filters of 66.5 (quality factor of attenuation Q=0.5) and 38.5 (Q=1) Hz, along with a 60 Hz centered band stop filter (Q=5), were used to screen out the high frequency unwanted noise.
For channel 2, which monitored the GSR signal, DC input coupling along with the most restrictive low pass hardware filter of 1 KHz was initially chosen to allow for the low frequency GSR signal to pass. After analysis of a sample signal frequency content, AC coupling along with a 0.05 Hz high pass hardware filter that minimized baseline drift, was used to allow for a higher resolution. To screen out unwanted high-frequency noise, heavy low and band pass filtering was employed (two low pass filters at 66.5 and 38.5 Hz and a band stop filter centered at 60 Hz, Q=5).
The sampling rate used to prevent signal aliasing was 200 samples/sec for both the ECG and the GSR signal. The gain for the ECG and GSR signals was set at 2000 and 5000, respectively. After data acquisition, the fast signals of the GSR plots were digitally filtered using smoothing (200 data points).
A total of nine subjects were instructed to drink 1L of water to induce urination urge and were connected to the BioPac ECG and GSR probes once they experienced an extreme urge to urinate. Three ECG electrodes were placed on the right-hand wrist and above each inner ankle. The electrode lead set (SS2L) connected to channel 1 of the BioPac acquisition unit was attached to the three electrodes. The GSR Transducers (SS3L) were connected to channel 2. Electrode gel was placed on the transducers, which were strapped on the index and middle fingers of the left hand. The subject was seated comfortably and instructed to keep eyes closed and not concentrate on extraneous distractions in order to reduce fluctuations in ECG and GSR caused by other sources of sympathetic stimulation. Two 30-second readings of ECG and GSR were taken successively. After recording the treatment (before urination) data, the subject was allowed to urinate. Following urination, the subject rested for 5 minutes, and measurements were taken in a similar fashion as used above and were used as the control data.
4. Results
4.1. ECG data
Values for each subject’s changing heart rates were obtained from the available tools of the BioPac system (e.g. ECG R-R bmp channel). T testing (t=1.96 at over 6000 data points and degrees of freedom for p=0.05) was used to compare each subject’s average heart rate data before and after urination (Table 1). The validating condition imposed was |X1-X2|>t*(SD12/N1+ SD22/N2) 0.5, where N1, X1, SD1 and N2, X2, SD2 are the sample size, average heart rate, and standard deviation before and after urination, respectively. In seven out of nine cases it was found that heart rate significantly increases under urination urge.
T testing for eight degrees of freedom (t=2.306 at p=0.05 level) was used to compare the average heart rate of the entire population before and after urination. The 95% confidence interval for the individual differences in average heart rates was calculated to be 1.68 to 12.17 bpm. Since zero was not included in this confidence interval, it was concluded that the difference in average heart rate is significantly positive. Thus, average heart rate was found to have increased from 69.93 bpm to 76.84 bpm under urination urge.
Table 1. Heart rate (bpm) averages with standard deviations before (treatment) and after (control) urination
Subject No. / Treatment /Control
/ Difference1 / 86.87 / 68.65 / 18.22
2 / 70.69 / 71.74 / -1.05
3 / 83.6 / 80.28 / 3.32
4 / 79.87 / 65.9 / 13.97
5 / 68.93 / 66.89 / 2.04
6 / 83.27 / 70.14 / 13.13
7 / 81.16 / 81.27 / -0.11
8 / 56.89 / 52.52 / 4.37
9 / 80.25 / 71.98 / 8.27
Average / 76.84 / 69.93 / 6.91
Stdev / 9.523 / 8.467 / 6.837
Variance of the individual and overall average heart rates was studied using Excel F-testing. In seven out of nine subjects, F values exceeded the critical values needed for significance at the 5% level. This shows that it is probable that the variability in an individual’s heart rate is larger when the urge to urinate is present (Figure 1).
Figure 1. Heart rate variability under treatment and control conditions for subject 1.
Variance for the overall average heart rate before and after urination of all individuals was evaluated, and the maximum critical F value of 1.61 necessary for confidence at the 5% level was also exceeded. It can be concluded that the heart rate varies significantly more among individuals under urination urge than in its absence. This shows that the heart is more easily excitable under a sympathetic stimulation such as the presence of the urge to urinate.
4.2. GSR data
Average GSR values and standard deviations for each individual recording (Table 2) were obtained from the BioPac calculation channels.
T testing (t=1.96 at over 6000 data points and degrees of freedom for p=0.05) was used similarly to the ECG analysis to evaluate whether or not urination urge increases, as expected, the average GSR. The results showed that only the average GSR of two subjects out of the nine total increased significantly with urination urge.
T testing for eight degrees of freedom (t=2.306 at p=0.05) was used in a similar manner to the ECG analysis to compare the average GSR values before and after urination for the entire subject population. The 95% confidence interval for the average differences between the GSR values before and after urination was calculated to be –1.71 to 0.13 Ohm-1. Since zero was included in the confidence interval, it was concluded that average GSR values do not change significantly when the urge to urinate is present.
F testing was used to assess the significance of individual and overall variance changes before and after urination. In two out of nine subjects, F values exceeded the critical values for approximately infinity degrees of freedom at the 5% confidence level. The change in the variances of the average GSR values pooled together for all nine subjects before and after urination was evaluated, and the critical F value of 2.48 (at maximum) at the 5% confidence level was not exceeded.
Table 2. Average GSR (Ohm-1) values and overall standard deviations before (treatment) and after (control) urination
Subject No. /Control
/ Treatment / Difference1 / -0.34 / 0.14 / 0.48
2 / 0.05 / -0.81 / -0.86
3 / -0.02 / -0.69 / -0.67
4 / 0.13 / 0.18 / 0.06
5 / -0.57 / 1.85 / 2.42
6 / -0.99 / -0.81 / 0.18
7 / 1.15 / -0.61 / -1.76
8 / -0.56 / -0.92 / -0.36
9 / -0.16 / 0.04 / 0.20
Average / -0.15 / -0.18 / -0.04
Stdev / 0.603 / 0.879 / 1.149
These two variance analyses lead to the conclusion that there is no significant difference between the variability of the GSR signal before and after urination.
5. Discussion
The manner in which each subject’s heart rate was determined was subject to reading error. Heart rate was obtained manually by simply highlighting the ECG peak-to-peak time intervals and using the corresponding calculation channel. Averaging many readings when computing the average heart rates and variances mitigates the repercussions of this type of error. As results show, individual heart rate variability is significantly increased under the urination urge in seven out of nine individuals, which makes the relative contribution of such reading errors to the overall result insignificant. A larger number of subjects as well as increased accuracy in measuring heart rates should insure that such results are more representative of the entire population.
Of particular concern were subjects under medication, which can alter the heart rate. The two subjects whose data did not show increased heart rate variability were under medication that increased their average heart rate. It can therefore be speculated that since their average heart rates were higher than their usual values even in the control case, they would display decreased variability under treatment. This could be due to a permanent and, presumably, stronger effect of the medicine on heart activity than the transient one induced by the sympathetic nervous system. Further study on different groups of subjects sorted out by the presence and nature of their medication should show whether or not increased heart rate variability under urination urge depends on the usage and nature of the respective medicine.
A source of possible experimental errors in obtaining the GSR values is the placement of the sensors on the skin. After the initial set of measurements for ECG and GSR, the subjects had to remove the GSR transducers from their fingers in order to urinate. Removal of the sensors from the fingers may have resulted in inconsistencies in GSR readings. The transducers measure the potential difference between two exact points on the skin, assuming that they do not move. When the sensors were taken off and reapplied to the skin, they measured GSR readings of a different set of conditions, because it is likely that they were not placed on the same initial locations on the fingers.
These errors in GSR precision are unlikely the cause for the insignificant changes in the average value and variance of this data. Since perspiration occurs in response to most sympathetic impulses, the effect of urination urge might be obscured by other sympathetic events that cannot be controlled in such an experiment.
While average heart rate and its variability could be used in applications that automatically signal urination urges in the disabled, GSR appears unsuitable as an indicator of such an urge.
6. References
[1] Campbell, Neil A., Biology, 4th Ed, The Benjamin/Cummings Publishing, Menlo Park, 1996.
[2] Lewis, D.G. The Analysis of Variance, Manchester University Press, Manchester, 1971.
[3] Pflanzer, Richard. BioPac Student Lab Neurological. BIOPAC Systems, Inc., Santa Barbara, 1999.