A time ~ (24~h)~
j A time w 1803 1
New context test
Passive Avoidance. Fig. 1. Experimental sequences for passive avoidance experiments (a) and contextual fear conditioning (b) with training and retention tests with information on the duration of subintervals as used in previous studies. Blue arrows (a) indicate when the instrumental response occurs, i.e., the active choice of transferring from the bright into the dark compartment. In contrast, in fear conditioning the fear response is commonly quantified on the basis of visually observed or computer-derived freezing as index of active suppression of ongoing behavior. The test in the new context serves as control to determine the specificity of the fear response elicited by the distinct stimuli provided in the conditioning context. US = unconditioned stimulus (foot shock); comp. = compartment. (Modified from Ogren et al. 2008.)
Step-Through Passive Avoidance (Inhibitory Avoidance)
The step-through task is a one-trial emotional memory task combining fear conditioning with an instrumental response, e.g., the active choice of an animal to avoid entering the dark compartment associated with an aversive event (Ogren 1985). The passive avoidance task differs from the typical fear conditioning experiment in which training and testing occur in a one-compartment box. The rodent placed in the apparatus is exposed to one or several foot shocks (shock intensities 0.5-0.8 mA). Memory retention is measured as the degree of suppression of motor behavior quantified on the basis of ► freezing.
The typical step-through passive avoidance test is conducted in a two-compartment box with one bright and one dark compartment connected by a sliding door. The subject is placed in the bright compartment and will after a defined time interval gain access to the dark compartment. When entering the dark compartment (training latency), the door will be closed and the subject will be subjected to a brief aversive stimulus (US; foot shock intensities in mice around 0.2-0.4 mA) that will lead to the formation of an association of the dark compartment with the US. For the retention test (usually 24 h after training) the animal is returned in the bright compartment with the sliding door open. The animal has now the option to avoid or enter the dark compartment by discriminating the bright (safe) from the dark (unsafe) compartment. The rapid acquisition of not making a response indicates that the test involves learned inhibition rather than loss of an innate response tendency. Usually one trial, i.e., one exposure to the inescapable shock is sufficient to suppress the innate preference of the rodents for the dark chamber of the apparatus. The acquisition of passive avoidance is measured either as a significant increase of the step-through latency compared to training latencies or as the decrease of the time spent by the subjects inside the dark chamber. Thus, the passive avoidance procedure has the advantage of simplicity in that both the safe and the noxious compartment are clearly defined as well as the correct adaptive response that is to refrain from entering the dark compartment.
In the step-down passive avoidance task the subject is placed on an elevated platform from which it can step down onto the floor below, mainly a shock grid. When stepping onto the floor, the animal will be subjected to one or several ► aversive stimuli (US; foot shock) that will lead to the formation of an association between the shock and the context resulting in avoidance or delay to step down when returned to the platform in the retention test. Unlike the step-through passive avoidance procedure, the retention test is performed in the original (punished)
context similar to fear conditioning. This feature may lead to a substantial amount of contextual freezing, which is rarely observed in the step-through task. Another important issue relates to the compartment size, i.e., the space available for exploration in a one-compartment box. It is recommended to offer sufficient space for exploration instead of having systems with minimal compartment size from which animals instantly transfer when moving. Step-down passive avoidance tasks are more sensitive to changes in locomotor activity than step-through passive avoidance tasks particularly if the platform is small. Drug-induced or lesion-induced alterations in locomotor activity may confound the actual avoidance response.
Since the majority of passive avoidance investigations are based on the use of step-through avoidance paradigms, the following sections will focus on this task.
In the available literature there is quite a range of different procedures and experimental designs. Procedural differences have considerable consequences for passive avoidance performance and variability in results. Therefore, important aspects of different procedures are briefly summarized in Table 1. Different passive avoidance systems exist that are customized or available from different vendors. It is important that they fulfill certain hardware/ software requirements for flexible use ideally with convenient software controls as indicated in Table 1.
Some pre-experimental procedures have considerable effects on the results. Handling by the experimenter has been shown to reduce variations between animals in the passive avoidance task probably by lowering the adverse effects of acute stress (Madjid et al. 2006). Another procedural modification, which reduces variations in results, is to allow the animal to explore both compartments for a short time (e.g., 1 min) 1 h before the actual passive avoidance training. However, this procedure may result in ► latent inhibition, e.g., reduced transfer latency, if the pre-exposure to the dark compartment is too long, if it occurs 24 h before training or if the strain of mice or rats is particularly sensitive to pre-exposure (Baarendse etal. 2008).
The most critical parameters in passive avoidance are the shock intensity and its duration. The avoidance response shows a very steep increase in the transfer latency with increasing shock intensity (Fig. 2). Thus, even small variations in intensity can change passive avoidance performance. Therefore, it is important to determine shock thresholds in all experiments, since large species and strain variations exist. To be able to study both enhancement and impairment of passive avoidance memory in the same experiment, different shock intensities can be used.
To enable demonstration of facilitatory effects on drugs on passive avoidance retention, a mildly aversive electrical current can be used while a strong current can be employed to explicitly study impairment of passive avoidance retention. This procedure provides a sufficiently wide for impairment and blockade of impairment when combining agonist and antagonist treatment (Fig. 3). An alternative approach is to use very long training latencies, e.g., up to 900 s, to have a sufficient dynamical range, or to combine it with pre-exposure (latent inhibition).
In addition, a wide range of retention measures is generally observed as a sign of large inter-individual variability. Therefore, larger group sizes (generally n =12—16/ group) are required to achieve required statistical power unless profound drug effects are observed. Statistical analysis may have to be performed with nonparametric analysis of variance (ANOVA) using the Kruskal-Wallis test followed by the Mann-Whitney U test for pair-wise comparisons due to deviations from normal distribution of individual data and large group variances or responses exceeding maximum transfer latencies (cut-off latencies).
Handling of animals just after training can reduce the retention latency considerably. Particularly in mice, it is important that they can remain in the dark compartment of the passive avoidance apparatus for at least 30 s before being removed and transferred to their holding or home cage. Thus, the animal has to be given sufficient time to process the training context. In fear conditioning, exposure to the shock instantly after placement in the conditioning context leads to the "immediate shock deficit'' in rats and mice, i.e., the absence of contextual conditioning. Thus, contextual information requires sufficient processing time to form an aversive association indicating the importance of the temporal relation of CS and US for associative learning. This is achieved by the initial confinement to the bright compartment. Another important rationale for the 30-s delay after US exposure is to avoid that mice form an aversive association of the US with the handling procedure (removal from the compartment after the US exposure).
The nature of contextual stimuli which are critical for associative learning in the passive avoidance are not well characterized. It seems likely that passive avoidance, similar to contextual fear conditioning, depends on multisen-sory associations rather than unisensory associations. The relative role of sensory cues representing the CS is not known at present. Sensory afferents are required for the
Compartment sizes (larger compartments allow for movements irrespective of lack of transfer) Door size and visibility (larger doors promote faster transfer during training)
Light intensity differences between compartments (should be profound, e.g., 10-100-fold difference) Door mechanism (a guillotine door falling on a mouse that is not yet completely in the dark compartment may act as US)
Shocker with sufficient sensitivity (ideally >0.1 mA increments) and range (e.g., 0-1 mA)
Housing (single or group) and transfer (short vs. long) to the experimental room (arousal affecting the actual training and testing)
Pre-exposure(s) before training, timing of pre-exposure (to reduce data variability), risk of latent inhibition
Handling (to reduce variability in training latencies)
Timing of drug treatment (pre- vs. post-training, pretest, state-dependent learning)
Timing of retention testing (from short-term [1 h] to long-term memory [24-? h]; remote memory
[>7 days] is hardly explored)
Test duration (cut-off time: e.g., 300 s vs. 600 s)
US intensity range used: low for facilitation vs. high for impairment
Current: constant vs. scrambled shock (generally scrambled shocks are more effective)
Extinction tests based on repetitive testing (with/without forced exposure to the dark compartment)*
Waiting/delay time before access to the dark compartment is provided (e.g., 60 s during training,
15 s during testing)*
Delay between dark compartment entry and US exposure to avoid escaping to the bright compartment*
Transfer latencies detection (full transition when the animal is in the dark compartment with all four feet)
Transitions between compartments (detection based on center of gravity may record transitions in mice during stretch-attend postures while exploring the door; this is not ideal for automated analysis)
Total time spent per compartment
US response assessment (important to avoid misinterpretations based on altered nociception) Activity measurements (some hypoactive and neophobic mouse strains such as A/J are unsuited for passive avoidance experiments despite implying to be good learners; therefore it is sometimes recommendable to include non-shocked controls)
Time of testing (fluctuations in passive avoidance response throughout the circadian cycle have been reported in rats)
Group sizes (normally n = 8/group is minimum, ideally n = 12-16/group)
Parametric (ANOVA) and nonparametric data analyzes (Kruskal-Wallis and Mann-Whitney U test) (depending on the normal distribution of individual data [group variance] or skewed data because of maximum latency cut-off)
*These parameters have not been systematically investigated
*These parameters have not been systematically investigated detection of the different stimuli provided in the passive avoidance test environment from tactile to visual cues, nociceptive pain receptors for US detection and possibly also olfactory cues. Besides processing in the thalamus (except for olfactory information), higher brain centers are involved for ► encoding, ► consolidation, and ► extinction. The ► amygdala, the ► hippocampus, and the various cortical areas are part of the neural network that subserve passive avoidance learning (Baarendse et al. 2008; Burwell et al. 2004; McGaugh 2004; Ogren et al. 2008).
A number of studies have shown that passive avoidance depends on hippocampal function and its ► NMDA-receptors. Thus, infusion of the NMDA-receptor antagonists, AP5, into the dorsal hippocampus of mice profoundly impairs passive avoidance retention (Baarendse et al. 2008). Also neurotoxic lesions of the corticohippocampal circuitry (perirhinal, postrhinal, and entorhinal cortex) cause profound deficits in passive avoidance learning (Burwell et al. 2004). These results indicate that the passive avoidance task requires processing of spatial information about the test environment.
Passive Avoidance. Fig. 2. Passive avoidance performance in male C57BL/6J mice as a function of US intensity using multiple measures. US responses (activity) were measured during training and compared to baseline activity (a). In the retention test, transfer latencies (b), the total time spent in the bright compartment (c), and number of entries into the dark compartment (d) were analyzed as a function of US intensity (US duration: 2 s). The basal exploratory activity in the initial 180-s period of exploration in the light compartment is indicated as dashed horizontal line (dotted lines: ±SEM) in panel a. The dashed horizontal line (dotted lines: ±SEM) in panel b indicates the mean training latency. Dotted lines at 300 s in panels b and c denote the cut-off time in the retention test. Error bars indicate SEM. comp. = compartment; *p < 0.05 and ***p <0.001 vs. 0.0 mA control group (based on Mann-Whitney U test for panel b).
The passive avoidance task differs from spatial tests such as the ► Morris water maze both with regard to sensitivity to parahippocampal lesions and modulation of cholinergic mechanisms, suggesting that these two spatial tasks differ in their dependence on processing of polymodal sensory information (Burwell et al. 2004; Ogren et al. 2008). Finally, it should be noted that mice with hippo-campal dysfunction may acquire passive avoidance, although suboptimally, using a hippocampus-independent strategy (see Baarendse et al. 2008).
A large body of literature points at the crucial role of the amygdala, a heavily innervated assembly of many different subnuclei that are essential for aversive learning (LeDoux 2000; McGaugh 2004). Among these subnuclei, the basolateral nucleus plays a major role in the convergence of CS and US for associative learning. The basolat-eral nucleus of the amygdala is connected to the central nucleus of the amygdala. Outputs from the central nucleus of the amygdala are essentially responsible for the expression of fear responses. This network triggers behavioral adjustments indicative of learning and memory through motor control and concomitant autonomic and endocrine adjustments via efferent pathways. The efferent pathways are specific for certain response domains but subserve both learned and innate responses. These brain areas include hypothalamic and various brain stem areas that will not be discussed here.
Application of Passive Avoidance in Neuropsychopharmacology
Passive avoidance is one of the most frequently used animal tests for studying learning and memory mechanisms and to identify compounds modifying cognitive processes. This task is often a first-line test in pharmaceutical companies. The task has also a prominent role in neuroscience research focusing on the role of neurotransmitters and molecular signals in learning and memory processes. An accumulated body of work has characterized the major multiple neurochemical systems and some of its molecular components which mediate or modulate this type of learning (Ogren et al. 2008). Research already in the 1970-1980s showed a significant role of cholinergic transmission in storage and retrieval of information. Thus, the nonselective muscarinic receptor antagonists such as ► scopolamine, when injected prior to passive avoidance training, were found to cause a dose-dependent impairment of passive avoidance retention (Bartus et al. 1982). Subsequent studies showed that administration of cholinomimetic drugs such as the acetylcholinesterase inhibitor ► physostigmine could partially antagonize the deficit caused by scopolamine. These findings formed part of the cholinergic hypothesis of geriatric memory dysfunction (Bartus et al. 1982) that provided the rationale for introducing ► cholinesterase inhibitors as anti-dementia drugs. A number of studies have also demonstrated a significant role for brain serotonin in passive avoidance learning. Both increases and decreases in brain 5-HT transmission have resulted in passive avoidance deficits ((Ogren et al. 2008) probably reflecting the involvement of multiple 5-HT receptors in this task (Misane and (Ogren 2000). However, through the use of 5-HT receptor subtype-specific ligands, such as selective 5-HT1A receptor agonists and antagonists, the role of 5-HT in passive avoidance has been more thoroughly investigated (see Fig. 3). With passive avoidance as the major behavioral test it has been possible to distinguish the modulatory action of pre- and post-synaptic 5-HT1A receptors in cognitive function ((Ogren et al. 2008). Based on these findings, the 5-HT1A receptors have emerged as an important target for drugs acting in psychopathologies characterized by mood disorders and disturbances in emotional memory ((Ogren et al. 2008).
A major problem in the analysis of drug effects in passive avoidance is that drug-induced changes in emotionality, motivation, or brain chemistry may be part of the training context. ► State-dependent learning refers to the situations in which retention is poor for rodents when the drug state during learning and retention differs. This means that information acquired under the drug state can only be retrieved when the animal is in the same drug state. Analyses of state-dependent learning can be an important tool by which to assess nonspecific effects on performance from associative learning mechanisms. The research design in passive avoidance
Passive Avoidance. Fig. 3. Passive avoidance performance based on transfer latencies in the retention test in male NMRI mice as a function of drug treatment using different shock (US) intensities during training as indicated. The full 5-HT1A agonist 8-OH-DPAT (a) facilitates at the low-dose range (0.01-0.03 mg/kg) and impairs at high-dose range (>0.1 mg/kg). The selective 5-HT1A antagonist NAD-299 (b) facilitates passive avoidance retention latencies in the dose range of 0.3-2 mg/kg. The muscarinic acetylcholine receptor antagonist scopolamine (c) impairs passive avoidance retention at 0.1 and 0.3 mg/kg. US duration was 1 s. Retention latency cut-off was 300 s. All drugs were injected either subcutaneously 15 min (a and b) or 40 min before training (c). Error bars indicate SEM; *p < 0.05 and **p < 0.01 vs. 0.0 mg/kg control group. (Modified from Madjid et al. 2006).
experiments should always include a test group receiving the same treatment at both training and test.
Advantages and Limitations of Passive Avoidance Experiments
The advantage with the passive avoidance procedure is that it is a single-trial task and it produces similar results in both mice and rats. Unlike multitrial tasks, single-trial tasks allow for a precise timing of drug injections either before or after training, or before the retention test. It is, therefore, possible to dissect out the possible contributions and effects of drug interventions on encoding, consolidation, and retention (memory retrieval and expression). Since the retention tests can be performed at variable time intervals after training, ► short-term (STM) and long-term memory (LTM) can be assessed separately. However, since subsequent tests are confounded by prior experiences, an animal can only be tested once. Post-training administration provides the opportunity for studying the endogenous modulation of memory consolidation.
There are also drawbacks with this one-trial task, including large inter-subject variability (discussed above) and different sensitivity to the shock exposures. The memory-strength (retention) is influenced by many factors including the mental state at the time oflearning, emotional reactivity and responses to stress hormones (McGaugh 2004). Experimental manipulations that cause changes in locomotion, pain perception or anxiety states can also confound measures of memory. Information on these behavioral domains is important to rule out alternative explanations rather than altered cognitive function(s).
Attenuated pain perception by pharmacological or genetic manipulations with analgesic effects may render a low US ineffective as punisher, thereby preventing passive avoidance learning. Knowledge about the US responsiveness (see Baarendse et al. 2008) is crucial to rule out passive avoidance impairments because of lack of US salience/perception. Since the US is given in the dark compartment, quantification is hardly possible unless the passive avoidance system used can detect the activity of experimental animals in the dark compartment. It is useful to install photo beams that transmit through the black walls or using infrared cameras for video tracking. At least the US reactions (e.g., vocalization and jumping) of the animals should be visually observed and noted by the experimenter to exclude a potential lack of perception.
There is often a general assumption that enhanced anxiety-like behavior is associated with increased fear (avoidance) responses. Although this cannot be ruled out, there is at this moment no clear evidence that this assumption is valid (Madjid et al. 2006). On the contrary, studies using different strains of mice show that the DBA/2J strain which are more anxious than the C57BL/6J strain, show reduced passive avoidance learning compared to C57BL/6J mice (Baarendse et al. 2008). In conclusion, despite their limitations, passive avoidance paradigms have turned out to be very useful in analyzing the neuro-chemical and molecular basis of cognition.
Passive avoidance tests have been used in research for more than 6 decades. The success of this behavioral test is probably based on the fact that it exploits adaptive behavioral mechanisms critical for survival and shaped through evolution. Moreover, the passive avoidance task is very flexible and allows for investigation of a large number of cognitive domains in neuropsychopharmacol-ogy and related disciplines. Yet, the full potential of this task in studying memory mechanisms is still not always well perceived in the research community. The advent of advanced passive avoidance systems will allow a much more penetrating analysis of domains of cognition with relevance for psychopathology such as latent inhibition and extinction of passive avoidance responses. In addition, the integration ofadditional measures such as neural responses and autonomic adjustments with passive avoidance will open up new avenues for an integrative analysis of avoidance behavior Therefore, it is suggested that the passive avoidance task has an important place in both current and future neuroscience research.
Was this article helpful?