Using Accelerometers to Determine the Cessation of Activity of Broilers

2007 Poultry Science Association, Inc. Using Accelerometers to Determine the Cessation of Activity of Broilers M. D. Dawson, M. E. Lombardi, E. R. Benson, 1 R. L. Alphin, and G. W. Malone Delaware Experimental Station, College of Agriculture and Natural Resources, University of Delaware, Newark 19716 Primary Audience: Complex Managers, Veterinarians, Researchers SUMMARY In recent years, the threat of a pandemic outbreak of avian influenza has become a global issue. The existence of zoonotic strains, causing cases of human infection throughout Eastern Asia, has increased interest in the prevention and containment of avian influenza outbreaks. Research has led to the development of a mass emergency depopulation method utilizing water-based foam for turkeys and chickens. Birds undergoing foam depopulation cannot be easily observed to determine time to death. As a result, a means to monitor birds and to determine time to death was needed to be developed in parallel to the new depopulation method. In this study, the use of an accelerometer for determination of the cessation of reflex reactions during the procedure was validated by comparing muscular cessation on external limbs to cardio-relaxation detected by an electrocardiogram. Electrocardiogram and accelerometer readings were taken from trials in which birds were euthanized by cervical dislocation and compared with other experiments in which water-based foam and CO 2 -polyethylene tent treatments were used. Statistical analysis indicated that accelerometers are valid sensors for the detection of the cessation of terminal convulsions in broilers. Key words: accelerometer, broiler, depopulation, electrocardiogram, foam, carbon dioxide polyethylene tent procedure 2007 J. Appl. Poult. Res. 16:583 591 doi:10.3382/japr.2007-00023 DESCRIPTION OF PROBLEM The spread of avian influenza (AI) poses a serious threat to poultry populations and a potential threat to humans. An outbreak of highpathogenic AI virus, H5N1 virus, has caused many human fatalities and has required the destruction of millions of birds [1]. The United States has been fortunate in that most domestic outbreaks have been of the low-pathogenic AI virus (LPAIV) that has less of a zoonotic potential. Most recent disease outbreaks in the United States have been quickly contained. The United States has suffered high-pathogenic AI virus outbreaks such as the 1983 Pennsylvania H5N2 outbreak in which 17 million birds were destroyed at a cost of US$61 million [2, 3]. More recently, an outbreak of LPAIV in Delaware and Maryland resulted in a ban on the import of US poultry in 30 nations in 2004 [4, 5]. The economic effect of AI outbreaks can be devastating, particularly in heavy poultry production areas such as Sussex County, Delaware, the highest poultry-producing county in the United States, where the 2004 outbreak re- 1 Corresponding author: ebenson@udel.edu

584 sulted in the destruction of 85,000 birds on 2 Delaware farms. For this reason, a new technique to rapidly contain and depopulate infected flocks was developed using water-based fire-fighting foam. One measure of effectiveness of a depopulation treatment is the elapsed time from treatment to death. During the early evaluation of the water-based foam depopulation technique, measurement of the elapsed time to death was not practical, because terminal movements could not be observed with birds fully immersed in opaque water-based firefighting foam. Monitoring of heart activity by a stethoscope required removal of 1 or more birds from the treatment to perform the observation, introducing an observer effect. The need to quantitatively monitor the depopulation process is a relatively unusual requirement associated with the development of a depopulation procedure, not a requirement for field use during emergency depopulation. To compensate for the difficulty in viewing the birds immersed in foam, the use of an accelerometer to measure terminal activity was evaluated. An accelerometer measures changes in velocity or speed. The accelerometer was used to measure vibration caused by a bird in terminal convulsions, where the observation of a flat line on monitoring equipment would indicate cessation of the terminal convulsive activity. In this study, the use of an accelerometer for monitoring the cessation of activity in broilers was evaluated. Background The CO 2 -polyethylene tent procedure used in the 2004 LPAIV outbreaks in Delaware was the control procedure in early experiments to determine the effectiveness of foam as a depopulation method [6]. In the CO 2 -polyethylene tent procedure, the birds are covered with polyethylene, and the volume of air under the polyethylene sheet is replaced with CO 2 gas. When subjected to high levels of CO 2, birds are anesthetized and then begin terminal convulsions before dying of hypoxia. Cessation time in CO 2 - polyethylene tent trials was declared by a poultry science expert when the cessation of terminal convulsions was observed. Five minutes after the induction of gas, a random bird was selected and checked for a heartbeat. JAPR: Research Report The CO 2 -polyethylene tent procedure used in Delaware was similar to the techniques used in the 2002 Virginia AI virus outbreak. Depopulation was handled in Virginia using 2 controlled atmosphere methods. The first method involved constructing a ground panel enclosure. A rectangular plywood enclosure was built at 1 end of a turkey house and covered by a tarp supported by a zigzag rope truss to prevent sagging [7]. Turkeys were herded into the enclosure in batches of 5,500 birds, and an average of 6 min 20 s were required for audible signs of activity to cease. In a second procedure for caged birds, a metal enclosure was placed over birds in live-haul cages on a flatbed truck. The second technique also required the euthanasia of birds in batches with the cessation of audible signs of activity occurring at an average of 1 min 28 s for 375 chickens. Detailed physiological studies were conducted in the United Kingdom, where the effects of CO 2 stunning of hens on electroencephalogram (EEG) suppression and loss of somatosensory-evoked potentials (SEP) were analyzed [8, 9]. In these studies, several physiological phases were studied, and data were provided for suppression of EEG, loss of SEP, period of convulsing (e.g., onset and duration of clonic and tonic phases), and finally EEG silence. In terms of the period of convulsions, the clonic phase is characterized by episodes of wing-flapping, whereas the tonic phase involved the birds becoming rigid and showing final paddling motions with their legs and wings extended [8]. In the first study, Raj et al. [8] stunned 17 hens in 45% CO 2. Raj et al. [8] found EEG suppression (unconsciousness) occurred in 21 ± 4 s, loss of SEP in 30 ± 2 s, and EEG silence (brain death) in 101 ± 18 s. In terms of the reflex response brought on by the suppression of brain activity, the onset of the clonic phase occurred at 45 ± 4 s and typically lasted 15 ± 5 s. The birds would typically enter the tonic phase at 68 ± 10 s for a duration of 25 ± 10 s. Using the data presented, visible movement typically ceased in 93 s. In a second study, 12 hens were euthanized using 30% CO 2 in Ar with 5% residual O 2 [9]. The CO 2 -O 2 gas mixture caused EEG suppression to occur in an average of 14.0 s, loss of

DAWSON ET AL.: DETERMINING BROILER CESSATION ACTIVITY 585 SEP in 17.1 s, and EEG silence in 58.0 s. Cessation of visible movement occurred on average in 60.0 s. The clonic phase typically set in at 17.1 s and lasted 14.6 s. The tonic phase typically started at 35.0 s. During early foam depopulation studies [6, 10], small polycarbonate euthanasia chambers were used. After treatment with foam, involuntary convulsions in birds would agitate and cause motion of the foam. In these experiments, the cessation of foam agitation was the best possible indication of the cessation of bird activity. Once foam agitation ceased, a poultry expert would call the cessation time. At 5 min after the introduction of foam, a random bird would be removed from the chamber and examined for a heartbeat. In all but 1 instance, all birds were found to have no detectable heartbeat within 5 min, and no bird had a detectable heartbeat at 10 min [6]. During depopulation experiments, it is not practical to examine many birds to determine cessation, cessation time, or both. Examination of individual birds is time-consuming and requires removing the birds from treatment. The use of EEG in large experiments is impractical, because surgery is required to implant the sensor. Although the use of an electrocardiogram (ECG) does not require surgery, it does require extended preparation time including bird preparation and attachment of multiple electrodes, making it impractical in the field. Because measurement of brain activity (EEG) or heart activity (ECG) is not practical under field conditions, an alternate method of monitoring the birds during controlled depopulation studies was needed. The use of an accelerometer, a device that can detect sudden acceleration changes, was suggested for detecting the time associated with the irreversible convulsions that accompany the loss of brain activity. Using an accelerometer provides several advantages over EEG and ECG. The preparation time is negligible, because it only requires securing the sensor to a limb (Figure 1). Accelerometers also require less equipment, because multiple sensors can be monitored from a single computer. Both EEG and ECG have high bandwidth requirements that restrict the number of birds that can be simultaneously monitored. The AVMA [11] considers cardiac arrest to be clinical death. Because there is little information on the detection of time of death through measurement of terminal movements, an analytical method was required to determine when death occurs in relation to the cessation of the convulsive period. Cervical dislocation, or the dislocation of the neck vertebrae from the cranium, is an approved method of euthanasia in poultry [11]. Cervical dislocation causes separation of the brain from the spinal cord and carotid arteries. Terminal movements including wing motion may be observed for several minutes in birds subjected to cervical dislocation. The observed spastic movements are reflex reactions resulting from sudden termination of signals from the brain and represent the irreversible clonic and tonic phases of convulsion preceding death. Based on the findings of Raj et al. [8, 9], brain death occurs at or shortly after the convulsive phase. Because the accelerometer senses motion and can detect the end of the convulsive phase, it can therefore serve as a relative indicator of the time of brain death. MATERIALS AND METHODS In a preliminary experiment (data not shown) to evaluate the best placement for an accelerometer, 3 locations (e.g., the wing, neck, and leg) of broilers were tested as possible contact points. Given that the cessation of final movements would serve as a critical time, the extremity that provided the longest time until a flat line was observed on monitoring equipment was desired. Accelerometer data indicated that the leg position provided the longest cessation times when a broiler was euthanized by cervical dislocation. After cessation of leg movement, motion could still be observed in the breast of the bird, but no practical means of attaching the accelerometer to detect chest motion was available. Although the heart will beat for several minutes after brain death in a bird, an accelerometer does not directly measure heart activity. To monitor heart activity, ECG electrodes were attached to the birds. Although the application of electrode pads was a stressful experience for the birds, ECG measurements showed that birds returned to a steady heartbeat within

586 JAPR: Research Report Figure 1. Accelerometers were attached to the leg of broilers to measure cessation of activity. seconds of the completion of preparation for ECG measurements. For this study, data were collected from a controlled experiment utilizing euthanasia by cervical dislocation and from 2 prior experiments using mass emergency depopulation methods. Experiment 1 describes the base cervical dislocation experiment conducted specifically for this study. Experiments 2 and 3 were conducted previously to investigate the following: 1. water-based fire-fighting foam as an effective method for the mass depopulation of floor-reared poultry in emergency scenarios, 2. stress experienced by the birds when subjected to different depopulation treatments, and 3. performance of foam-generation equipment under field conditions. Accelerometer readings were collected during experiments 2 and 3, but those experiments were not conducted specifically to evaluate the use of accelerometers. To validate the use of accelerometers in determining the cessation of convulsions, a stepwise statistical analysis was conducted. The first step was to determine if the cessation times for external extremities measured by the accelerometer were significantly different from heart relaxation times as observed via an ECG when broilers are euthanized by cervical dislocation. The second step was to compare data gathered from laboratory-based foam depopulation trials to the cessation-relaxation times from the cervical dislocation experiment. The final stage of analysis was to compare CO 2 -polyethylene tent depopulation and foam depopulation data collected under field-simulated conditions to the cervical dislocation data set.

DAWSON ET AL.: DETERMINING BROILER CESSATION ACTIVITY 587 In all 3 experiments, 1 of 2 PCB Piezotronics [12] shear mode accelerometers, models 353B16 and 352C66, were used on each bird as shown in Figure 1. The 353B16 was the initial accelerometer, with a sensitivity of 1.02 mv/(m/s 2 ) ± 10% (10 mv/g ± 10%) capable of operating over a range of ±4,905 m/s 2 of peak (±500 g of peak). The second higher-sensitivity accelerometer, 10.2 mv/(m/s 2 ) ± 10% (100 mv/ g ± 10%), which was introduced in later trials, had an operational range of ±491 m/s 2 of peak (±50 g of peak). For the purpose of the depopulation study, the signal characteristic of interest was the time at which a flat line begins, which minimized any difference in accelerometer characteristics. The accelerometer output was passed through a PCB Piezotronics [12] single-channel signal conditioner, model 480C02, and recorded independently using National Instruments [13] PCI-6036E data acquisition card. The monitoring interface was a custom-written virtual instrument developed in National Instruments [13] LabVIEW data acquisition and analysis software. The cessation time of interest was the period from the beginning of convulsive activity until detectable motion ceases. Each bird was also instrumented with ECG sensors in the first and second experiment. By design, no ECG data were collected in the field during the third experiment. Each bird had ECG monitoring pads secured onto their left leg, right wing, and right leg. The ECG output was recorded on a BIOPAC Systems Inc. [14] MP30A acquisition unit using BSL Pro monitoring software. Due to bandwidth requirements, ECG monitoring and virtual instrumentation software for the accelerometer were operated on separate computers. The time at which the heart relaxes, observed as a stable low-amplitude heart signal with a decreasing beat rate, was the signal characteristic of interest. Bird activity in each experiment was recorded over a 300-s period for both ECG and accelerometer sensors. All testing was performed under the approval and guidelines of the University of Delaware Agricultural Animal Care and Use Committee and followed the guidelines laid out by the Federation of Animal Science Societies [15]. Experiment 1 Twelve randomly selected 6-wk-old broilers were instrumented with both accelerometer and ECG sensors. One bird per trial was euthanized via cervical dislocation. Each broiler was placed in a 113-L (30 gal) chamber to restrict the range of movement during clonic convulsions. Cervical dislocation was applied 7 s after sensor recordings began. Experiment 2 A laboratory study was conducted to determine the differences between depopulation by the introduction of fire-fighting foam and the CO 2 -polyethylene tent procedure. The purpose of the experiment was to determine the physiological cause of death in birds subjected to each treatment and to measure the corticosterone hormone levels before and after treatment to gauge stress. Ten broilers (1-bird replicates) were subjected to 3 depopulation treatments: foam enriched with CO 2, foam without CO 2, and the CO 2 -polyethylene tent procedure. A total of 30 broilers were tested. Further detail of experiment 2 is discussed in Benson et al. [10]. For the evaluation of the accelerometer, missing data were omitted, and six 1-bird replicates, each of the foam with CO 2 and foam without CO 2, were used for a total of 12 foam depopulation observations. It was previously determined that no significant difference existed between foam with and without CO 2 [10], so both foam treatments can be handled as a single foam treatment. For the foam trials, each bird was placed into 1 of 2 prefilled 113-L (30 gal) chambers. A solution of 160 ml of Ansul [16] Jet-X high expansion foam concentrate and 6 L of tap water were agitated to create foam. Sensor recording began at the moment the bird was introduced to the foam chamber. Inadequate accelerometer data were collected for the CO 2 -polyethylene tent replicates precluding comparisons to the CO 2 -polyethylene tent procedure from this experiment. Experiment 3 A study was conducted to compare the effects of foam depopulation against the CO 2 - polyethylene tent procedure under simulated field conditions. Experiment 3 was also used to

588 JAPR: Research Report Table 1. Descriptive statistics for accelerometer-detected cessation times and electrocardiogram (ECG)-detected cardiac-relaxation times for each treatment across all experiments Cessation and relaxation Experiment Treatment Sensor Mean (s) SD (s) Min (s) Max (s) 1 Cervical dislocation Accelerometer 128 51 53 219 1 Cervical dislocation ECG 154 49 80 219 2 Foam Accelerometer 48 23 15 85 2 Foam ECG 64 17 37 93 2 Foam Accelerometer 109 57 61 269 3 Poly tent 1 Accelerometer 61 16 28 98 1 CO 2 -polyethylene tent procedure. evaluate the expected performance of a prototype depopulation foam generator by comparing the reliability of properly (good) and improperly generated (bad) foam. Of the 2 foam formulations, good foam was found to have a very high probability of depopulation success; bad foam had a lower likelihood of depopulation success, in which success was defined as a 0% survival rate. For the purposes of this study, only good foam and CO 2 -polyethylene tent data were analyzed. For the CO 2 -polyethylene tent procedure, each bird was instrumented and placed in a 79 65 79 cm (31 25.5 31 in.) clear polycarbonate euthanasia chamber. The CO 2 gas was discharged from a gas cylinder for 60 s via a hose entering the chamber from the top. The hose was secured to the bottom of the chamber to prevent hose movement. The chamber was lined with clear polyethylene while an excess polyethylene sheet was then folded over the birds and held in place with weights. The birds were not in a hermetically sealed environment, and CO 2 gas could escape during the depopulation procedure. The rate of the introduction of CO 2 gas was much higher than the rate of escape during the induction period. Also, because CO 2 is heavier than air, much of the escaping air during the induction period would be displaced breathable atmo- sphere. Sensor recording started simultaneously with the opening of the valve on the CO 2 cylinder. The foaming trials were conducted in a field open to the elements. The birds were placed in an open triangular enclosure made of two 1.22 2.44 m (4 8 ft) plywood boards. The third wall was a section of plywood cut to a height of approximately 0.61 m (2 ft) so that foam could be introduced into the enclosure from a cart-mounted foam generator. Each bird was instrumented, placed in the enclosure, and then foam was introduced into the enclosure until it overflowed the front panel. Foam generation was performed using a prototype Kifco [17] Avi-FoamGuard foam generator system. Analysis Statistical analysis was conducted in SAS [18]. The data collected in each experiment were generally nonnormal, requiring analysis using nonparametric statistical tests techniques. The Wilcoxon signed rank test was used for analysis of sensor differences. Comparison of depopulation treatments was analyzed with the exact Wilcoxon 2-sample test. All tests were conducted at the 5% significance level (α = 0.05). Table 2. Results of the Wilcoxon signed rank test performed on differences between cessation times measured by the accelerometer and cardiac relaxation detected by electrocardiogram (ECG) Mean, SD Median Signed P-value Experiment 1 d (s) (s) (s) rank S P(= S ) 1 26 29 26 29 0.0195 2 22 16 23 22.5 0.0039

DAWSON ET AL.: DETERMINING BROILER CESSATION ACTIVITY 589 Table 3. Results of the Wilcoxon 2-sample exact test indicating the significance of differences between treatments Observations, Sum of Expected under SD under Mean One-sided Experiment Sensor Treatment 1 n scores S H 0 E 0 (S) H 0 σ 0 (S) score a P-value 2 Accelerometer Cervical dislocation 2 12 213.0 150.0 17.31 17.75 <0.0001 2 ECG Foam 2 9 46.0 99.0 14.07 5.11 <0.0001 3 Accelerometer Poly tent 3 15 169.0 247.5 26.46 11.27 0.0011 3 Accelerometer Cervical dislocation 4 12 206.5 180.0 22.57 17.21 0.1246 3 Accelerometer Cervical dislocation 5 12 244.0 168.0 20.48 20.33 <0.0001 1 The Wilcoxon exact test sums the scores of the observations of the smaller set of samples. 2 Experiment 2 compared foam depopulation to cervical dislocation. 3 Foam depopulation vs. the CO 2 -polyethylene tent procedure. 4 Foam depopulation vs. cervical dislocation. 5 CO 2 -polyethylene tent procedure vs. cervical dislocation. RESULTS AND DISCUSSION The accelerometer and ECG were used to measure cessation of activity for 3 experiments. Table 1 shows the descriptive statistics for each experiment. The results of the Wilcoxon tests are summarized in Tables 2 and 3. Experiment 1 In the cervical dislocation experiment, 24 data points were collected for a total of 12 observations (n = 12) per sensor. A 1-way analysis was conducted on the calculated differences between the cessation times measured by the accelerometer and the cardiac relaxation times recorded by the ECG. The mean difference ( accelerometer ECG ) used in the signed rank test was 26 ± 29 s (Table 2). The hypothesis for the sensor comparison was: H 0 : accelerometer ECG = 0 [1] H A : accelerometer ECG 0 [2] The P-value for the signed rank test was 0.0195, indicating significance (P 0.05). Therefore, a significant difference exists between the cessation of (observable) movement and the relaxation of the heart. Experiment 2 Three statistical tests were conducted on the results from the second experiment. The first test was similar to the test conducted in experiment 1 (Table 2). A total of 21 observations were collected from the second experiment (n accelerometer = 12, n ECG = 9, accelerometer ECG = 22 ± 16 s) and analyzed. Using the same hypothesis as in experiment 1 (equations [1] and [2]), the signed rank test returned a P-value of 0.0039, indicating significant differences between the times recorded by the accelerometer and ECG consistent with the findings for experiment 1. Because the accelerometer and ECG measure different times, treatment comparisons had to be performed on a per-sensor basis. The hypothesis for comparing the foam data to the cervical dislocation data from experiment 1 is: H 0 : foam = CD [3] H A : foam CD [4] Separate tests were conducted for the data acquired from the accelerometer (n accelerometer, foam = 12, n accelerometer, CD = 12) and ECG (n ECG, foam = 9, n ECG, CD = 12). For both the accelerometer and ECG, the 1-sided exact Wilcoxon 2- sample test returned P < 0.001 (Table 3). Therefore, both the accelerometer and ECG see foam depopulation and cervical dislocation as different treatments. Experiment 3 Field conditions for experiment 3 made it impractical to collect ECG measurements. As a result, heart relaxation time was not measured, and no sensor comparisons could be made. Three Wilcoxon 2-sample tests were conducted to compare treatments. The first test compared the data collected from foam depopulation (n = 17) and CO 2 -polyethylene tent (n = 15) repli-

590 cates. The hypothesis is as shown in equations [5] and [6]. H 0 : foam = poly [5] H A : foam poly [6] The Wilcoxon test returned P = 0.0011, indicating that foam depopulation has different cessation times from the CO 2 -polyethylene tent procedure (Table 3). Thus, comparisons to the base treatment, cervical dislocation (experiment 1), must be performed separately. When the CO 2 -polyethylene tent procedure was compared with cervical dislocation, the 2 treatments were found to have significantly different cessation times, but no significant difference was found between foam and cervical dislocation (Table 3). The latter finding is inconsistent with the findings from experiment 2. DISCUSSION Using the UK study [8, 9] as a reference, birds lose consciousness before entering into involuntary convulsions due to the suppression of brain activity. The loss of SEP also occurs at or before the onset of the clonic phase of terminal seizures, meaning that the birds are not responsive to external stimuli once convulsions begin. Cervical dislocation, which involves severing the vertebrae, has the same physiological effects with EEG suppression, loss of SEP and onset of convulsions occurring almost instantaneously. The CO 2 -polyethylene tent procedure is very similar to the stunning techniques used by Raj et al. [8, 9], indicating that as the result of EEG suppression and loss of SEP (unconsciousness), the birds enter into terminal convulsions. During the convulsive phase, there is a significant difference between the time that the heart relaxes and the time that cessation of terminal convulsions occur. It is understood that heart activity will continue for up to several minutes after brain death. Using combined statistical data from Raj et al. [8, 9], brain death occurs at approximately 81 s. Cessation times as detected by the accelerometer ranged from 25 to 179 s, whereas EEG suspension occurred between 58 to 119 s. Therefore, the birds are JAPR: Research Report effectively brain dead at the point at which the convulsive phase ends. The AVMA defines clinical death in animals as cardiac arrest, but the study of poultry physiology shows that death occurs in phases. First, brain activity is suppressed, and then response to external stimuli ceases (i.e., loss of SEP). Convulsions occur once brain activity is irreversibly suppressed. Brain death occurs at or shortly after the cessation of the convulsive phase. Finally, cardiac arrest in birds subjected to cervical dislocation, the CO 2 -polyethylene tent procedure, and foam treatment always occurs at some time after the 300-s recording session, well outside the range of EEG suspension times found by Raj et al. [8, 9]. Because the ECG detects cardiac muscle reflexes and the accelerometer measures gross body movement, ECG cessation times are therefore different from accelerometer cessation times. The Wilcoxon signed rank tests performed in experiments 1 and 2 confirm that the 2 sensors are not interchangeable. When the data collected in experiment 2 were compared with the data from experiment 1, the cessation (accelerometer) and cardio-relaxation (ECG) times for cervical dislocation and the foam treatment were found to be statistically different per sensor. Based on the mean cessation and relaxation times (Table 1), it can be stated with 95% confidence that cessation of observable activity, the conclusion of the tonic phase of convulsions, occurs before cardiac relaxation and subsequent cardiac arrest. In experiment 3, only accelerometer data were collected, and a significant difference was detected between the foam and CO 2 -polyethylene tent treatments. The differences between the foam and CO 2 stunning found here are consistent with the findings in Benson et al. [10]. One inconsistency was found in that no significant difference was found by the Wilcoxon 2- sample exact test between the cessation times for foam and cervical dislocation as in experiment 2. The inconsistency may be attributed to the differences in the quality of foam produced in the laboratory vs. foam generated by the prototype foam generator in the field. Unpublished findings based on experiment 3 did indicate that foam quality can have a significant

DAWSON ET AL.: DETERMINING BROILER CESSATION ACTIVITY 591 effect on the effectiveness of foam depopulation (data not shown). Cessation times measured by the accelerometer occur at the end of the tonic phase of convulsions and before heart relaxation. Based on the findings by Raj et al. [8, 9], brain death occurs at or shortly after the end of the convulsive phase. Therefore, accelerometers can be used to determine the end of the convulsive phase and as an estimator of the time of brain death. A further study including the use of EEG will be required to determine the mean time difference between the cessation times detected by the accelerometer and the actual time that brain death occurs. CONCLUSIONS AND APPLICATIONS 1. Cessation of activity, as measured by the accelerometer, occurs at or about the same time as brain death. 2. Cardiac relaxation typically occurs after the convulsive phase. 3. Cardiac arrest typically did not occur during the 5-min observation period. 4. Ideally, in a similar future study, 3 depopulation-euthanasia procedures including cervical dislocation, CO 2 -polyethylene tent, and foam treatments should be conducted recording simultaneous ECG, EEG, and accelerometer observations. The depopulation treatments used should be conducted under both controlled laboratory as well as uncontrolled field-simulation conditions. REFERENCES AND NOTES 1. World Health Organization. 2007. Cumulative Number of Confirmed Human Cases of Avian Influenza A/(H5N1) Reported to WHO. http://www.who.int/csr/disease/avian_influenza/country/ cases_table_2007_02_03/en/index.html Accessed Feb. 2007. 2. Goldblatt, J., C. Barrish, and B. Tadesse. 2004. So far, no more Delaware farms test positive. News Journal (Wilmington, DE) 13:1, 2A. 3. Lu, H., P. A. Dunn, E. A. Wallner-Pendleton, D. J. Henzler, D. C. Kradel, J. Liu, D. P. Shaw, and P. Miller. 2004. Investigation of two H7N2 avian influenza outbreaks in two broiler breeder flocks in Pennsylvania, 2001-02. Avian Dis. 48:26 33. 4. Capua, I., and D. J. Alexander. 2004. Avian influenza: Recent developments. Avian Pathol. 33:393 404. 5. Montgomery, J. 2004. Bird flu in Texas a threat. News Journal (Wilmington, DE) 25:1, 2A. 6. Dawson, M. D., E. R. Benson, G. W. Malone, R. L. Alphin, I. Estevez, and G. L. Van Wicklen. 2006. Evaluation of foam-based mass depopulation methodology for floor-reared meat-type poultry operations. Appl. Eng. Agric. 22:787 793. 7. Kingston, S. K., C. A. Dussalt, R. S. Zaaidlicz, N. H. Faltas, M. E. Geib, S. Taylor, T. Holt, and B. A. Porter-Spalding. 2005. Evaluation of two methods for mass emergency depopulation of poultry in disease outbreaks. J. Am. Vet. Med. Assoc. 227:730 738. 8. Raj, A. B. M., N. G. Gregory, and S. B. Wotton. 1990. Effect of carbon dioxide stunning on somatosensory evoked potentials in hens. Res. Vet. Sci. 49:355 359. 9. Raj, A. B. M., S. B. Wotton, and P. E. Whittington. 1992. Changes in the spontaneous and evoked electrical activity in the brain of hens during stunning with 30 percent carbon dioxide in argon with 5 percent residual oxygen. Res. Vet. Sci. 53:126 129. 10. Benson, E. R., G. W. Malone, R. L. Alphin, M. D. Dawson, C. R. Pope, and G. L. Van Wicklen. 2007. Foam-based mass emergency depopulation of floor-reared meat-type poultry operations. Poult. Sci. 86:219 224. 11. AVMA. 2001. Report of the AVMA panel on depopulation. J. Am. Vet. Med. Assoc. 218:669 698. 12. PCB Piezotronics Inc., Depew, NY. 13. National Instruments Corporation, Austin, TX. 14. BIOPAC Systems Inc., Goleta, CA. 15. Federation for Animal Science Societies. 1999. Guidelines for the Care and Use of Agricultural Animals in Agricultural Research and Teaching. 1st rev. ed. Fed. Anim. Sci. Soc., Savoy, IL. 16. Ansul Inc., Marinette, WI. 17. Kifco Inc., Havana, IL. 18. SAS Institute Inc., Cary, NC. Acknowledgments This research was supported by University of Delaware College of Agriculture and Natural Resources, US Poultry and Egg Association, USDA-Veterinary Services, and Kifco Inc. We would like to acknowledge the contributions of V. Lariccia, C. R. Pope, G. L. Van Wicklen, K. Johnson, J. Kelly, and S. L. Collier.