- Email: [email protected]
Computer analysis and quantification of periaqueductal grey-induced defence behaviour
Journal of Neuroscience Methods 58 (1995) 157-161
uter analysis and quantification of periaqueductal defence behaviour S. Beckett *, CA. Marsden Department
of Physiology
and Pharmacology, Received
30 May
University
Medical
1994; revised
School Nottingham,
14 September
Queens Medical
1994; Accepted
15 September
Centre,
Nottingham
AlG7 UG,
UK
1994
Abstract The pharmacological study of midbrain-evoked aversive behaviours is commonly measured in terms of electrical escape thresholds, although studies which examine the actual expression of the behaviour are being introduced. A computer-driven automated tracking system has been developed to record and analyse the very rapid locomotor activity produced by the defence response. Stimulation of the midbrain periaqueductal grey (PAG) matter in rats with the excitatory amino acid o,L-homocysteic acid (DLH) produced very rapid and short-lasting escape behaviour which was tracked by computer. The system, VideoTrack, determines the speed and distance travelled by the animal and stores and statistically analyses the data. VideoTrack was able to detect a pharmacological modifications of the response; anti-aversive effects produced by intra-PAG pretreatment with the 5-HT,, agonist 8hydroxy 2-(di-n-propylamino)tetralin (S-OHDPAT) and its reversal by peripheral pretreatment with the SHT,, antagonist N-[2-[-fZ-methoxyphenyl)-l-piperazinyl]ethyl]-N-(pyridinyl)cyclohexanecarboxanide (WAY 100635). fiywords:
Periaqueductal grey; Defence response; Aversion; 8-OHDPAT;
This paper describes a multitask computer-driven automated procedure for recording and analysing data obtained by inducing escape behaviour in the rat. The system allows rapid overt and intense behaviours, such as defence, to be automatically monitored and analysed in detail. Defence behaviour occurs naturally as a protection mechanism allowing an organism to respond to unconditioned stimuli in the appropriate way. It consists of a behavioural escape component, supported by cardiovascular and analgesic changes (Fernandez de Molina and Hunsperger, 1962; Blanchard et al., 1987) similar to those observed in human panic and anxiety disorder (Deakin and Graeff, 1991). The defence response in the rat can be produced by electrical stimulation of central areas such as the amygdala, hypothalamus and midbrain periqaueductal grey (PAG) which comprise e main neural substrate of fear or anxiety (Graeff,
* Corresponding 709259.
author.
Tel.:
(44-602)
0165.0270/95/$09..50 0 1995 Elsevier SSDI 0165-0270(94)00170-7
Science
709413;
B.V. All
Fax:
(44-602)
rights
resewed
WAY 100635; SHT,,;
Computer tracking
19901. Specific chemical stimulation techniques using excitatory amino acids such as D,L-homocysteic acid (DLH) (Hilton and Redfern, 1986; Beckett et al., 1992) have highlighted the dorsal region of the PAG as an area particularly rich in cell bodies involved in this aversive behaviour. Several neurotransmitter systems have been associated with PAG-mediated aversion including GARA and benzodiazepines, excitatory amino acids, opiates and 5-HT (for review see Graeff et al., 19%). Serotonin, however, has recently received attention, due to its connection with anxiety states and reduced PAG serotonergic function is associated with pro-aversion whereas enhanced function is associated with antiaversion (for review see Graeff, 1990). Such results, however, only describe threshold effects an little about the expression of aversive behaviour itself. By stimulating the PAG directly using DLII the full behavioural repertoire can be observed, quantified and drug-induced changes in expression detected (Reckett et al., 1992). Rats placed in a circular arena are given a microinjection (250 nl) of DLII into the dorsal PAG, stimulating defence behaviour. The duration of the response together with the number of jumps and arena
158
S. Beckett,
CA.
Marsden
/Journal
of Neuroscience
Methods
58 (1995)
157-161
revoiutions made then provide a measure of the magnitude or intensity of the response. Comparisons of behavioural responses prior to and following drug pretreatment allow changes in the response to be detected. This process is, however, very time consuming and relatively subjective. In addition it reveals little about the behavioural profile of the response across time. The aim of the present study was to assess the ability of a computerised tracking system based on differential frame analysis to detect and measure the very fast behaviours associated with defence and detect pharmacological modification of the response. The results of the computer-based measurements were also compared with those made manually by an observer.
can be divided into zones which can be considered individually or grouped together. Experimental measures scored during an experiment are totally user definable. For each piece of apparatus the user can specify global and zone specific measures as well as complex locomotor events covering a number of zones, allowing a flexible system configuration to be achieved which will measure a wide variety of locomotor activities. More subtle behaviours? or ones which cannot be tracked for any reason, can be measured by the operator via a number of user defined keys. In addition to the comprehensive and flexible tracking facilities the system provides a test regime, randomising animals and treatments if required. All the tracking data stored by the computer and can be replayed and re analysed at any time allowing modifications in the apparatus or experiment set up to be explored.
2.1. Description of the computer system
2.2. Experimental procedure
VideoTrack has been designed and developed in collaboration with CPL systems (Cambridge, UK) to track a single animal around a user-defined piece of apparatus and record a range of behaviours performed over a wide range of speeds. The program runs under windows on IBM compatible PCs employing a standard video camera and digitising board. Experiments are either run on-line, i.e., directly from the camera input or retrospectively from video recordings. Tracking of the animal is achieved by the computer calculating the ‘difference’ between 2 sequential pictures (typical operating speeds are between 10 and 15 frames/s). If an animal is the only moving object in the picture, then any difference detected is attributable to the animal. Thus the system can be used on any piece of apparatus where the animal is clearly visible. Several software facilities also allow the tracking procedure to be adapted specifically to cope with particular pieces of apparatus, which because of inherent problems within the test environment would be difficult to monitor. Water based apparatus such as the Morris water maze, are provided with a facility to filter out moving reflections which interfere with the tracking. Measurement of escape behaviour is extremely difficult to achieve due to the explosive nature of the response and the high locomotor speeds which the animals achieve. This problem has been overcome by expanding the sampling window to accommodate rapid movements allowing the system to constantly track the animal without losing the image. The differential nature of this method of tracking allows accurate measurements to be made. Some alternative tracking methods rely on measurements made by tracking the point of highest contrast but this falls to monitor rapid locomotor behaviours. The system allows the user to define the apparatus regardless of its size or shape. Each piece of apparatus
DLH stimulation and local drug administration were performed via an indwelling intracerebral PAG cannula through which a microinjection could be made, the details of which are detailed elsewhere (Beckett et al., 1992). Briefly, cannulae aimed at the dorsal PAG (AP-6.7, ML-1.7, DV-S.lmm relative to bregma, Paxinos and Watson, 1982), were stereotaxically implanted under halothane anaesthesia. The wound was sutchered and the animal allowed at least 7 days to recover before behavioural testing. On day 1 a stainless steel microinjector, loaded with DLH (5 nmol in 250nl), was inserted into the guide. 50 i saline
40
+ X-QHDPAT
WAY100635 8.OHDPAT+
z
+ DLH
c DLH
E
drug treatment Fig. 1. Attenuation of DLH (5 nmol in 250 nl) induced defence response duration by intra-PAG pretreatment with the S-NIT,, agonist COHDPAT (10 nmol in 250 nl) and reversal by peripheral pretreatment with the SHT,, antagonist WAY 100635 (0.1 mg/kg, S.C.). * * P < 0.01 vs. DLH alone; ‘P < 0.05 vs. 8-OHDPAT pretreatment, l-way ANOVA Duncan’s new multiple range test; n = 11. Behaviour was monitored from video recordings by an observer.
S. Beckett, Table 1 Effect of 8-QHDPAT fence response
and WAY
100635
C.A. Marsden
on the DLH-induced 8-OHDPAT
Jumps DLH alone Saline S.C. + 8-OHDPAT S.C. + DLH WAY 100635 S.C. + B-OHDPAT SC. + DLH Revolutions DLH alone Saline S.C. + X-OHDPAT S.C. + DLH WAY 100635 S.C. + 8-OHDPAT S.C. + DLH
/Journal
of Neuroscience
de-
(10 nmol)
1.1+0.4 0.8 + 0.6 1.4 k 0.7 5.8 + 1.1 1.2kO.3 * 3.4 + 0.9
Attenuation of DLH (5 nmol in 250 nl) induced jumps and arena revolutions by intra-PAG pretreatment with the 5-HT,, agonist 8-OHDPAT (10 nmol in 250 nl) and reversal by peripheral pretreatantagonist WAY 100635 (0.1 mg/kg SC.). ment with the SHT,, * P < 0.05 vs. DLH alone. One-way ANOVA with post-hoc Duncan’s new multiple range test; n = 3 1). Behaviour was monitored from video recordings by an observer.
The animal was placed in a high walled circular test arena (75 cm@) and the injection performed over a 5 second period. The resultant behaviour was observed and recorded on video for 5 minutes. 48 hours later (day 3) animals were given a peripheral injection of saline (0.9% lml/kg s.c.) followed 20 minutes later by intra-PA6 8-OHDPAT (10 nmols in 250111s.c.) and 30 minutes later by DLH. On day 5 a peripheral injection of WAY100635 (O.lmg/kg SC.) was given 20 minutes prior to a second GOHDPAT injection (10nmols in 250111). This was followed 10 minutes later by a final stimulation. At the end of the experiment the animals were killed and the injection site marked with Pontamine Table 2 Effect of SOHDPAT and WAY 100635 fence response measured by the videotrack
on the DLH-induced system
Methods
58 (1995)
Blue dye (250 nl, 0.5%) before removing the brains and storing in saline prior to sectioning. Videotapes were analysed at the end of t ment, either (1) manually, in which case the duration of the response (time between onset and cessation of the explosive motor behaviours), the number of jumps and the number of arena revolutions made throughout the 5-min observation period were scored blind to the drug treatments as previously described (Bebkett et al., 1992) and matched later, or (2) by playing the tapes through the computer system. This produced values for the speed of the animal and distance travelled for the entire observation period and over continuous 20-s 0.92
0.80.7-
x
0.6-
--e-d -
t
DLHalone d+8-OHDP*~+IDLH wAYmx35+ S-OHDPAT+DLH
B 0.5$
0.4-
2
0.30.2" O.l-
0.0 ! 0
* 5 20 secondtime
10
15
10
15
10
1;
bins
de0
5
20secondtimebins
8-OHDPAT (10 nmol) Average speed (m/s) DLH alone Saline S.C. + &-OHDPAT S.C. + DLH WAY 100635 S.C. + 8-OHDPAT S.C. + DLH Maximum speed (m/s) DLH alone Saline S.C. + 8-OHDPAT S.C. + DLH WAY 100635 S.C. + 8-OHDPAT S.C. + DLH Distance travelled (m) DLH alone Saline S.C. + 8-OHDPAT S.C. + DLH 23.1 + 1.8 WAY 100635 S.C. + 8-OHDPAT S.C. + DLH
159
157-161
0.28 f 0.4 0.17kO.2 = 0.34 * 0.04 ++ 1.9 +2.2 1.58+0.16 1.84kO.13 29.7
il.1
5 20secondtimebins
29.8412.5
The effect of intra-PAG pretreatment with the .5-HT,, agonist 8-OHDPAT (10 nmol in 250 nl) and peripheral pretreatment with the S-HI’,, antagonist WAY 100635 (0.1 mg/kg s.c.) on (a) the average locomotor speed of the animal (m/s), (b) maximum speed (m/s) and distance travelled (m) produced by DLH (5 nmol in 250 nl) stimulation and measured throughout the 5-min observation period. * P < 0.05 vs. DLH alone; ‘P < 0.05; ‘+P < 0.01 vs. SOHDPAT pretreatment. One-way ANOVA with post-hoc Duncan’s new multiple range test; IZ = 11).
Fig. 2. The effect of intra-PAG pretreatment with the 5-HT,, agonist 8-OHDPAT (10 nmol in 250 nl) and peripheral. pretreatment with the SHT,, antagonist WAY 100635 (0.1 mg/kg, Y.c.) on (A) the average locomotor speed of the animal (m/s), (B) maximum speed (m/s) and (C) distance travelled (m) produced by DLH (5 nmol in 250 nl) stimulation and measured as 20-s time bins across the 5-min observation period. * P < 0.05; ’ )L P < 0.01 vs. DLH alone; ‘P < 0.05 vs. B-OHDPAT pretreatment, 2-way ANOVA Duncan’s new multiple range test; n = 11). Behavioural measurements derived from video analysis using the VideoTrack system.
160
time bins. The circular single zone.
S. Beckett,
CA.
Marsden
/Journal
arena was considered
of Neuroscience
as a
2.3. Data analysis
Data was analysed using l-way ANOVA with posthoc Duncan’s new multiple range test. Time bin measurements were analysed using a 2-way ANOVA. DLH (Sigma UK) 8-hydroxy 2-(di-n-propylamino) tetralin (8-OHDPAT, RBI UK), and N-[2-[-(2-methor;yphenyl)-l-piperazinyl]ethyl]-N-(pyridinyl)cyclohexa~ecarbo~anide (WAY 100635, Wyeth, UK) were dissolved in sterile physiological saline (0.9%, pH 7.4).
3.3.. Manual measurements
DLI-I injected into the dorsal PAG produced overt escape behaviours characteristic of the defence response including brisk running and jumping with occasional vocalisations. The response occurred 2-6 s after injection and lasted for 36 + 8 s (Fig. 1). Intra-PAG pretreatment with 8-OHDPAT 10 min prior to DLH stimulation resulted in a significant reduction in response duration and the number of arena revolutions
Methods
58 (1995)
1.57-161
(Fig. 1, Table 1). WAY 100635 given ~eri~~~~~~ly 20 min before 8-OHDPAT and 10 min prior to DLH stimulation significantly reversed the EbONDPATmediated attenuation. This was reflected by the response duration although the number of jumps and arena revolutions did not change significantly. 3.2. Computer measurements
Measurements of 3 different parameters were made: average speed (m/s>, maximum speed (m/s> and total distance travelled (ml. The results were expressed for the total 5 min of the observation period and as continuous 20-s time bins. Table 2 shows the results obtained for the total 5-min observation period. Each parameter was reduced by intra-PAG pretreatment with SOHDPAT although only average speed achieved significance. Similarly this effect could be reversed by peripheral WAY 100635 but only average speed showed a significant change. Data produced as time bins showed a significant reduction in average speed, m~im~rn speed and distance travelled within the first 40 s of the response, produced by SOHDPAT pretreatment. This was significantly reversed by WAY 100635 (Fig. 2). 8-OHDPAT-treated animals also showed a. significantly higher average and maximum speeds and dis-
images stored on video for computer display video signals digitised
i
measurement and analysis of differential frame sequences and results displayed
live picture Fig. 3. Schematic representation of the VideoTrack system operation: images obtained by a video camera, suspended above the arena, are captured by a digitiser. The system then analyses the sequence of digitised pictures for differences. Any difference detected is attributable to the movement of the animal. The data is analysed, stored to disk and a visual representation of the animals movements is displayed alongside a hve picture on the computer monitor. A video recorder connected between the camera and digitiser allows behaviours to be recorded for retrospective analysis.
S. Beckett,
CA.
Marsden
/Journal
of Neuroscience
tance values on or after the 10th time bin compared to animals treated with DLH alone. Fig. 3 shows a schematic representation of the system. 4. Discussion
Intra-PAG administration of DLH produced the characteristic escape behaviour, quantifiable in terms of the duration, number of jumps and arena revolutions. Measurements of speed and distance travelled provided additional descriptive data relating to the intensity of the response. Displaying the data as continuous 20-s time bins revealed the time course of the response. Animals displayed the most intense activity during the first 20-s time period. The highest average speeds, maximum speeds and distance travelled were recorded during the initial time period. These values fell substantially during the following 20-s time period dropping by over 50% in the case of distance travelled. By the 3rd time period, i.e., within 60 s of stimulation the values had stabilised, remaining at a steady level throughout the rest of the observation period although a tendency to increase could be seen around the 10th time period. The response decay is likely to occur through combined factors such as fatigue and removal or inactivation of DLH from around the injection site. Ten-minute intra-PAG pretreatment with 8OHDPAT produced a decrease in the observer-rated measures of defence. Response duration, number of arena revolutions and number of jumps were all significantly less compared to saline treatment. Computer analysis revealed that although the average 5-min values showed a decrease following 8-OHDPAT, only average speed reached significance. Time period analysis, however, revealed significant decreases in the animals average and maximum speeds and the distance travelled over the first two 20-s time bins. In addition it can be seen that the initial values obtained were reduced compared to saline reflecting not just a shorter response but a general lower level of intensity. These findings agree well with previous studies (Beckett et al., 1992; Graeff et al., 1993) which have demonstrated that PAG-associated aversion is under serotonergic control, mediated at least in part, by postsynaptic 5-HT,, receptors (for review see Deakin and Graeff, 1991; Graeff et al., 1993). This effect of S-OHDPAT could be reversed by peripheral administration of the selective 5-HT,, antagonist WAY 100635. Response duration and number of revolutions showed significant
Methods
58 (1995)
157-161
161
increases compared to S-OHDPAT pretreatment. Similarly the computer measurements showed an antagonistic effect of WAY 100635 on the 8-OHDPAT attenuation; however, only average speed values attained significance. This effect was highlighted by the time course measurements. The reduction in response parameters produced by 8-OHDPAT over the first 2 time periods was significantly reversed by WAY 100635, increasing both the duration and intensity of the response. Computer-driven tracking methods based on the differential sequence analysis of digitised images provides a valuable addition to the measurement of rapid movement allowing detailed non-subjective analysis of behavioural parameters previously difficult to obtain. The present study is the first to examine in detail the time course of stimulated defence behaviours in the rat and time bin analysis provides a highly sensitive measure of the effects of drugs on the response. Such a system may prove valuable for use not only in established behavioural procedures but also in refining and developing new ones.
Acknowle$gements
We thank Wyeth Pharmaceuticals for the gift of WAY 100635 and the Wellcome Trust for financial support.
References Beckett, S.R.G., Lawrence, A.J., Marsden, C.A. and Marshall, P.W. (1992) Attenuation of a chemically induced defence response by 5-HT, receptor agonists administered into the periaqueductal grey. Psychopharmacology, 108: 110-114. Deakin, J.F.W. and Graeff, F.G. (1991) 5-HT and mechanisms of defence. J. Psychopharmacol., 5: 305-315. Fernandez De Molina, A. and Hunsperger, R.W. (1962) Organisation of the subcortical system governing defence and flight reactions in the cat. J. Physiol. (Land)., 160: 200-213 Graeff, F.G. (1990) Brain defence mechanisms and anxiety. In: M. Roth, G.D. Burrows, and R. Noyles (Eds.), Handbook of Anxiety, Vol. 3, Elsevier, Amsterdam, pp. 307-354. Graeff, F.G., Silveira, M.C., Nogueira, R.L., Audi, E.A. and Oliveira, R.M. (1993) Role of amygdala and periaqueductal grey in anxiety and panic. Behav. Brain. Res., 58: 123-121. Hilton, SM. and Redfern, W.S. (1986) A search for brainstem cell groups integrating the defence reaction in the rat. J. Physiol., 378: 213-228. Paxinos, A. and Watson, C. (19821 The Rat Brain in Stereotaxic Coordinates, Academic Press, New York.
uter analysis and quantification of periaqueductal defence behaviour S. Beckett *, CA. Marsden Department
of Physiology
and Pharmacology, Received
30 May
University
Medical
1994; revised
School Nottingham,
14 September
Queens Medical
1994; Accepted
15 September
Centre,
Nottingham
AlG7 UG,
UK
1994
Abstract The pharmacological study of midbrain-evoked aversive behaviours is commonly measured in terms of electrical escape thresholds, although studies which examine the actual expression of the behaviour are being introduced. A computer-driven automated tracking system has been developed to record and analyse the very rapid locomotor activity produced by the defence response. Stimulation of the midbrain periaqueductal grey (PAG) matter in rats with the excitatory amino acid o,L-homocysteic acid (DLH) produced very rapid and short-lasting escape behaviour which was tracked by computer. The system, VideoTrack, determines the speed and distance travelled by the animal and stores and statistically analyses the data. VideoTrack was able to detect a pharmacological modifications of the response; anti-aversive effects produced by intra-PAG pretreatment with the 5-HT,, agonist 8hydroxy 2-(di-n-propylamino)tetralin (S-OHDPAT) and its reversal by peripheral pretreatment with the SHT,, antagonist N-[2-[-fZ-methoxyphenyl)-l-piperazinyl]ethyl]-N-(pyridinyl)cyclohexanecarboxanide (WAY 100635). fiywords:
Periaqueductal grey; Defence response; Aversion; 8-OHDPAT;
This paper describes a multitask computer-driven automated procedure for recording and analysing data obtained by inducing escape behaviour in the rat. The system allows rapid overt and intense behaviours, such as defence, to be automatically monitored and analysed in detail. Defence behaviour occurs naturally as a protection mechanism allowing an organism to respond to unconditioned stimuli in the appropriate way. It consists of a behavioural escape component, supported by cardiovascular and analgesic changes (Fernandez de Molina and Hunsperger, 1962; Blanchard et al., 1987) similar to those observed in human panic and anxiety disorder (Deakin and Graeff, 1991). The defence response in the rat can be produced by electrical stimulation of central areas such as the amygdala, hypothalamus and midbrain periqaueductal grey (PAG) which comprise e main neural substrate of fear or anxiety (Graeff,
* Corresponding 709259.
author.
Tel.:
(44-602)
0165.0270/95/$09..50 0 1995 Elsevier SSDI 0165-0270(94)00170-7
Science
709413;
B.V. All
Fax:
(44-602)
rights
resewed
WAY 100635; SHT,,;
Computer tracking
19901. Specific chemical stimulation techniques using excitatory amino acids such as D,L-homocysteic acid (DLH) (Hilton and Redfern, 1986; Beckett et al., 1992) have highlighted the dorsal region of the PAG as an area particularly rich in cell bodies involved in this aversive behaviour. Several neurotransmitter systems have been associated with PAG-mediated aversion including GARA and benzodiazepines, excitatory amino acids, opiates and 5-HT (for review see Graeff et al., 19%). Serotonin, however, has recently received attention, due to its connection with anxiety states and reduced PAG serotonergic function is associated with pro-aversion whereas enhanced function is associated with antiaversion (for review see Graeff, 1990). Such results, however, only describe threshold effects an little about the expression of aversive behaviour itself. By stimulating the PAG directly using DLII the full behavioural repertoire can be observed, quantified and drug-induced changes in expression detected (Reckett et al., 1992). Rats placed in a circular arena are given a microinjection (250 nl) of DLII into the dorsal PAG, stimulating defence behaviour. The duration of the response together with the number of jumps and arena
158
S. Beckett,
CA.
Marsden
/Journal
of Neuroscience
Methods
58 (1995)
157-161
revoiutions made then provide a measure of the magnitude or intensity of the response. Comparisons of behavioural responses prior to and following drug pretreatment allow changes in the response to be detected. This process is, however, very time consuming and relatively subjective. In addition it reveals little about the behavioural profile of the response across time. The aim of the present study was to assess the ability of a computerised tracking system based on differential frame analysis to detect and measure the very fast behaviours associated with defence and detect pharmacological modification of the response. The results of the computer-based measurements were also compared with those made manually by an observer.
can be divided into zones which can be considered individually or grouped together. Experimental measures scored during an experiment are totally user definable. For each piece of apparatus the user can specify global and zone specific measures as well as complex locomotor events covering a number of zones, allowing a flexible system configuration to be achieved which will measure a wide variety of locomotor activities. More subtle behaviours? or ones which cannot be tracked for any reason, can be measured by the operator via a number of user defined keys. In addition to the comprehensive and flexible tracking facilities the system provides a test regime, randomising animals and treatments if required. All the tracking data stored by the computer and can be replayed and re analysed at any time allowing modifications in the apparatus or experiment set up to be explored.
2.1. Description of the computer system
2.2. Experimental procedure
VideoTrack has been designed and developed in collaboration with CPL systems (Cambridge, UK) to track a single animal around a user-defined piece of apparatus and record a range of behaviours performed over a wide range of speeds. The program runs under windows on IBM compatible PCs employing a standard video camera and digitising board. Experiments are either run on-line, i.e., directly from the camera input or retrospectively from video recordings. Tracking of the animal is achieved by the computer calculating the ‘difference’ between 2 sequential pictures (typical operating speeds are between 10 and 15 frames/s). If an animal is the only moving object in the picture, then any difference detected is attributable to the animal. Thus the system can be used on any piece of apparatus where the animal is clearly visible. Several software facilities also allow the tracking procedure to be adapted specifically to cope with particular pieces of apparatus, which because of inherent problems within the test environment would be difficult to monitor. Water based apparatus such as the Morris water maze, are provided with a facility to filter out moving reflections which interfere with the tracking. Measurement of escape behaviour is extremely difficult to achieve due to the explosive nature of the response and the high locomotor speeds which the animals achieve. This problem has been overcome by expanding the sampling window to accommodate rapid movements allowing the system to constantly track the animal without losing the image. The differential nature of this method of tracking allows accurate measurements to be made. Some alternative tracking methods rely on measurements made by tracking the point of highest contrast but this falls to monitor rapid locomotor behaviours. The system allows the user to define the apparatus regardless of its size or shape. Each piece of apparatus
DLH stimulation and local drug administration were performed via an indwelling intracerebral PAG cannula through which a microinjection could be made, the details of which are detailed elsewhere (Beckett et al., 1992). Briefly, cannulae aimed at the dorsal PAG (AP-6.7, ML-1.7, DV-S.lmm relative to bregma, Paxinos and Watson, 1982), were stereotaxically implanted under halothane anaesthesia. The wound was sutchered and the animal allowed at least 7 days to recover before behavioural testing. On day 1 a stainless steel microinjector, loaded with DLH (5 nmol in 250nl), was inserted into the guide. 50 i saline
40
+ X-QHDPAT
WAY100635 8.OHDPAT+
z
+ DLH
c DLH
E
drug treatment Fig. 1. Attenuation of DLH (5 nmol in 250 nl) induced defence response duration by intra-PAG pretreatment with the S-NIT,, agonist COHDPAT (10 nmol in 250 nl) and reversal by peripheral pretreatment with the SHT,, antagonist WAY 100635 (0.1 mg/kg, S.C.). * * P < 0.01 vs. DLH alone; ‘P < 0.05 vs. 8-OHDPAT pretreatment, l-way ANOVA Duncan’s new multiple range test; n = 11. Behaviour was monitored from video recordings by an observer.
S. Beckett, Table 1 Effect of 8-QHDPAT fence response
and WAY
100635
C.A. Marsden
on the DLH-induced 8-OHDPAT
Jumps DLH alone Saline S.C. + 8-OHDPAT S.C. + DLH WAY 100635 S.C. + B-OHDPAT SC. + DLH Revolutions DLH alone Saline S.C. + X-OHDPAT S.C. + DLH WAY 100635 S.C. + 8-OHDPAT S.C. + DLH
/Journal
of Neuroscience
de-
(10 nmol)
1.1+0.4 0.8 + 0.6 1.4 k 0.7 5.8 + 1.1 1.2kO.3 * 3.4 + 0.9
Attenuation of DLH (5 nmol in 250 nl) induced jumps and arena revolutions by intra-PAG pretreatment with the 5-HT,, agonist 8-OHDPAT (10 nmol in 250 nl) and reversal by peripheral pretreatantagonist WAY 100635 (0.1 mg/kg SC.). ment with the SHT,, * P < 0.05 vs. DLH alone. One-way ANOVA with post-hoc Duncan’s new multiple range test; n = 3 1). Behaviour was monitored from video recordings by an observer.
The animal was placed in a high walled circular test arena (75 cm@) and the injection performed over a 5 second period. The resultant behaviour was observed and recorded on video for 5 minutes. 48 hours later (day 3) animals were given a peripheral injection of saline (0.9% lml/kg s.c.) followed 20 minutes later by intra-PA6 8-OHDPAT (10 nmols in 250111s.c.) and 30 minutes later by DLH. On day 5 a peripheral injection of WAY100635 (O.lmg/kg SC.) was given 20 minutes prior to a second GOHDPAT injection (10nmols in 250111). This was followed 10 minutes later by a final stimulation. At the end of the experiment the animals were killed and the injection site marked with Pontamine Table 2 Effect of SOHDPAT and WAY 100635 fence response measured by the videotrack
on the DLH-induced system
Methods
58 (1995)
Blue dye (250 nl, 0.5%) before removing the brains and storing in saline prior to sectioning. Videotapes were analysed at the end of t ment, either (1) manually, in which case the duration of the response (time between onset and cessation of the explosive motor behaviours), the number of jumps and the number of arena revolutions made throughout the 5-min observation period were scored blind to the drug treatments as previously described (Bebkett et al., 1992) and matched later, or (2) by playing the tapes through the computer system. This produced values for the speed of the animal and distance travelled for the entire observation period and over continuous 20-s 0.92
0.80.7-
x
0.6-
--e-d -
t
DLHalone d+8-OHDP*~+IDLH wAYmx35+ S-OHDPAT+DLH
B 0.5$
0.4-
2
0.30.2" O.l-
0.0 ! 0
* 5 20 secondtime
10
15
10
15
10
1;
bins
de0
5
20secondtimebins
8-OHDPAT (10 nmol) Average speed (m/s) DLH alone Saline S.C. + &-OHDPAT S.C. + DLH WAY 100635 S.C. + 8-OHDPAT S.C. + DLH Maximum speed (m/s) DLH alone Saline S.C. + 8-OHDPAT S.C. + DLH WAY 100635 S.C. + 8-OHDPAT S.C. + DLH Distance travelled (m) DLH alone Saline S.C. + 8-OHDPAT S.C. + DLH 23.1 + 1.8 WAY 100635 S.C. + 8-OHDPAT S.C. + DLH
159
157-161
0.28 f 0.4 0.17kO.2 = 0.34 * 0.04 ++ 1.9 +2.2 1.58+0.16 1.84kO.13 29.7
il.1
5 20secondtimebins
29.8412.5
The effect of intra-PAG pretreatment with the .5-HT,, agonist 8-OHDPAT (10 nmol in 250 nl) and peripheral pretreatment with the S-HI’,, antagonist WAY 100635 (0.1 mg/kg s.c.) on (a) the average locomotor speed of the animal (m/s), (b) maximum speed (m/s) and distance travelled (m) produced by DLH (5 nmol in 250 nl) stimulation and measured throughout the 5-min observation period. * P < 0.05 vs. DLH alone; ‘P < 0.05; ‘+P < 0.01 vs. SOHDPAT pretreatment. One-way ANOVA with post-hoc Duncan’s new multiple range test; IZ = 11).
Fig. 2. The effect of intra-PAG pretreatment with the 5-HT,, agonist 8-OHDPAT (10 nmol in 250 nl) and peripheral. pretreatment with the SHT,, antagonist WAY 100635 (0.1 mg/kg, Y.c.) on (A) the average locomotor speed of the animal (m/s), (B) maximum speed (m/s) and (C) distance travelled (m) produced by DLH (5 nmol in 250 nl) stimulation and measured as 20-s time bins across the 5-min observation period. * P < 0.05; ’ )L P < 0.01 vs. DLH alone; ‘P < 0.05 vs. B-OHDPAT pretreatment, 2-way ANOVA Duncan’s new multiple range test; n = 11). Behavioural measurements derived from video analysis using the VideoTrack system.
160
time bins. The circular single zone.
S. Beckett,
CA.
Marsden
/Journal
arena was considered
of Neuroscience
as a
2.3. Data analysis
Data was analysed using l-way ANOVA with posthoc Duncan’s new multiple range test. Time bin measurements were analysed using a 2-way ANOVA. DLH (Sigma UK) 8-hydroxy 2-(di-n-propylamino) tetralin (8-OHDPAT, RBI UK), and N-[2-[-(2-methor;yphenyl)-l-piperazinyl]ethyl]-N-(pyridinyl)cyclohexa~ecarbo~anide (WAY 100635, Wyeth, UK) were dissolved in sterile physiological saline (0.9%, pH 7.4).
3.3.. Manual measurements
DLI-I injected into the dorsal PAG produced overt escape behaviours characteristic of the defence response including brisk running and jumping with occasional vocalisations. The response occurred 2-6 s after injection and lasted for 36 + 8 s (Fig. 1). Intra-PAG pretreatment with 8-OHDPAT 10 min prior to DLH stimulation resulted in a significant reduction in response duration and the number of arena revolutions
Methods
58 (1995)
1.57-161
(Fig. 1, Table 1). WAY 100635 given ~eri~~~~~~ly 20 min before 8-OHDPAT and 10 min prior to DLH stimulation significantly reversed the EbONDPATmediated attenuation. This was reflected by the response duration although the number of jumps and arena revolutions did not change significantly. 3.2. Computer measurements
Measurements of 3 different parameters were made: average speed (m/s>, maximum speed (m/s> and total distance travelled (ml. The results were expressed for the total 5 min of the observation period and as continuous 20-s time bins. Table 2 shows the results obtained for the total 5-min observation period. Each parameter was reduced by intra-PAG pretreatment with SOHDPAT although only average speed achieved significance. Similarly this effect could be reversed by peripheral WAY 100635 but only average speed showed a significant change. Data produced as time bins showed a significant reduction in average speed, m~im~rn speed and distance travelled within the first 40 s of the response, produced by SOHDPAT pretreatment. This was significantly reversed by WAY 100635 (Fig. 2). 8-OHDPAT-treated animals also showed a. significantly higher average and maximum speeds and dis-
images stored on video for computer display video signals digitised
i
measurement and analysis of differential frame sequences and results displayed
live picture Fig. 3. Schematic representation of the VideoTrack system operation: images obtained by a video camera, suspended above the arena, are captured by a digitiser. The system then analyses the sequence of digitised pictures for differences. Any difference detected is attributable to the movement of the animal. The data is analysed, stored to disk and a visual representation of the animals movements is displayed alongside a hve picture on the computer monitor. A video recorder connected between the camera and digitiser allows behaviours to be recorded for retrospective analysis.
S. Beckett,
CA.
Marsden
/Journal
of Neuroscience
tance values on or after the 10th time bin compared to animals treated with DLH alone. Fig. 3 shows a schematic representation of the system. 4. Discussion
Intra-PAG administration of DLH produced the characteristic escape behaviour, quantifiable in terms of the duration, number of jumps and arena revolutions. Measurements of speed and distance travelled provided additional descriptive data relating to the intensity of the response. Displaying the data as continuous 20-s time bins revealed the time course of the response. Animals displayed the most intense activity during the first 20-s time period. The highest average speeds, maximum speeds and distance travelled were recorded during the initial time period. These values fell substantially during the following 20-s time period dropping by over 50% in the case of distance travelled. By the 3rd time period, i.e., within 60 s of stimulation the values had stabilised, remaining at a steady level throughout the rest of the observation period although a tendency to increase could be seen around the 10th time period. The response decay is likely to occur through combined factors such as fatigue and removal or inactivation of DLH from around the injection site. Ten-minute intra-PAG pretreatment with 8OHDPAT produced a decrease in the observer-rated measures of defence. Response duration, number of arena revolutions and number of jumps were all significantly less compared to saline treatment. Computer analysis revealed that although the average 5-min values showed a decrease following 8-OHDPAT, only average speed reached significance. Time period analysis, however, revealed significant decreases in the animals average and maximum speeds and the distance travelled over the first two 20-s time bins. In addition it can be seen that the initial values obtained were reduced compared to saline reflecting not just a shorter response but a general lower level of intensity. These findings agree well with previous studies (Beckett et al., 1992; Graeff et al., 1993) which have demonstrated that PAG-associated aversion is under serotonergic control, mediated at least in part, by postsynaptic 5-HT,, receptors (for review see Deakin and Graeff, 1991; Graeff et al., 1993). This effect of S-OHDPAT could be reversed by peripheral administration of the selective 5-HT,, antagonist WAY 100635. Response duration and number of revolutions showed significant
Methods
58 (1995)
157-161
161
increases compared to S-OHDPAT pretreatment. Similarly the computer measurements showed an antagonistic effect of WAY 100635 on the 8-OHDPAT attenuation; however, only average speed values attained significance. This effect was highlighted by the time course measurements. The reduction in response parameters produced by 8-OHDPAT over the first 2 time periods was significantly reversed by WAY 100635, increasing both the duration and intensity of the response. Computer-driven tracking methods based on the differential sequence analysis of digitised images provides a valuable addition to the measurement of rapid movement allowing detailed non-subjective analysis of behavioural parameters previously difficult to obtain. The present study is the first to examine in detail the time course of stimulated defence behaviours in the rat and time bin analysis provides a highly sensitive measure of the effects of drugs on the response. Such a system may prove valuable for use not only in established behavioural procedures but also in refining and developing new ones.
Acknowle$gements
We thank Wyeth Pharmaceuticals for the gift of WAY 100635 and the Wellcome Trust for financial support.
References Beckett, S.R.G., Lawrence, A.J., Marsden, C.A. and Marshall, P.W. (1992) Attenuation of a chemically induced defence response by 5-HT, receptor agonists administered into the periaqueductal grey. Psychopharmacology, 108: 110-114. Deakin, J.F.W. and Graeff, F.G. (1991) 5-HT and mechanisms of defence. J. Psychopharmacol., 5: 305-315. Fernandez De Molina, A. and Hunsperger, R.W. (1962) Organisation of the subcortical system governing defence and flight reactions in the cat. J. Physiol. (Land)., 160: 200-213 Graeff, F.G. (1990) Brain defence mechanisms and anxiety. In: M. Roth, G.D. Burrows, and R. Noyles (Eds.), Handbook of Anxiety, Vol. 3, Elsevier, Amsterdam, pp. 307-354. Graeff, F.G., Silveira, M.C., Nogueira, R.L., Audi, E.A. and Oliveira, R.M. (1993) Role of amygdala and periaqueductal grey in anxiety and panic. Behav. Brain. Res., 58: 123-121. Hilton, SM. and Redfern, W.S. (1986) A search for brainstem cell groups integrating the defence reaction in the rat. J. Physiol., 378: 213-228. Paxinos, A. and Watson, C. (19821 The Rat Brain in Stereotaxic Coordinates, Academic Press, New York.