Best "supraspinal" nociceptive assays?

Looking for some input here. I want to try CPA, but it seems not so easy to create any kind of consistent results.

Maybe thermal gradient? Formalin (licking)? Hot plate (jumping/licking)? Any suggestions for mechanical assays?

Any ideas would be great!

@ShanTan
See this thread:

We are now trying to set up the CPA in the model of Paclitaxel and SNI. I took a look in the Cheng paper. We are trying to avoid white and black boxes. We are using one white and one striped (black and white) chamber or one striped vertical and one horizontal. You should remember mice love dark places. It works well when we apply acetone. Another important technical details that is different from Cheng is that our chamber has one middle non-related chamber.
I will keep you posted and send additional information.

1 Like

Thanks Thiago! I tried the Cheng setup. I had a special chamber made. Truth is, the mice didn’t even like the dark very much more. It was still 50/50! I used dark red plexiglass so that I could still see inside but to the mice it should be dark. Light can still come up from the floor, but not so much.

Hope your experiments work well. Please keep us posted.

this is a subsection piece from one of our unpublished (hopefully soon) reviews, that very briefly mentions some behaviors we like:

" Measuring pain affect in preclinical models by linking neural activities to volitional behaviors

The brain is all about movement. Undeniably, brains are not functionally designed to represent the “truth” of the world. Neurons in the brain do not interact with the environment, but represent the relevant information to move the body, and keep the organism alive and propagating. Primordial species did not require perception; it was sufficient to move its mass of cells just a little to avoid dying. With evolutionary pressures driving the expanded complexity of brains for deploying future planning and decision-making processes, we might then observe the emergence of perception—new integrative neural circuit motifs that can directly compute past, present, and future states for inferential predictions that enable appropriate avoidance or approach actions. The Latin word “ emovere ,” where e - means ‘out’ and movere means ‘move’, forms the basis of the English word “emotion” and derives the term “motivation.” Thus, definitions of emotion and motivation are rooted in terms describing bodies in action. In this light, we can employ behavior as a readout of internal states, such as pain affect. However, it is a complex and difficult, but not impossible, task to assign a function to a specific neuron, circuit, or network for pain perception that is established from a behavioral readout.

In humans, experimental pain can be measured by simple self-reflective or self-evaluative reports, along with visual analog scales and detailed questionnaires. Recently, in attempt to identify visualized biomarkers of pain, machine-learning classifiers applied to fMRI and EEG datasets have decoded unique patterns of activity that correlate with noxious stimulus events or self-reports (Davis et al., 2020; Tayeb et al., 2020; Wager et al., 2013). However, basic pain neurobiology research has a clear need for better assays and endpoints of acute and chronic pain with superior predictive validity for translational target identification and compound screening (Mogil, 2019). A key issue with current assays, such as the von Frey mechanical sensitivity assay, is the experimenter themselves (Bove, 2006). The experimenter provides a subjective determination of a reflexive-paw withdrawal as a “pain-like” response or not. Critically, reflexes are not a primary endpoint for assessing human chronic pain; this severely restricts the translational utility of reflexive assays like von Frey. Equally, experimenter presence, i.e., a relatively large mammal next to or underneath the rodent, is itself a looming threat and a cue of imminent noxious stimuli. This repeated threat or stressor can greatly alter nociceptive processing via endogenous analgesia but is rarely addressed in data analysis (Hawranko et al., 1994).

Our group’s best attempt, thus far, to quantify affective-motivational behavior divides the cascade of stereotyped behaviors that immediately follow a noxious stimulus. In mice, brief noxious stimuli can elicit several distinct behavioral responses: 1. Withdrawal reflexes: rapid reflexive retraction or digit splaying of the paw, 2. Affective-motivational responses: directed licking and biting of the paw (termed ‘attending’), extended lifting or guarding of the paw, and/or escape responses characterized by hyperlocomotion, rearing or jumping away from the noxious stimulus. Paw withdrawal reflexes are classically measured in studies of hypersensitivity and involve simple spinal cord and brainstem circuits (as these behaviors are observed in decerebrated rodents). In contrast, affective-motivational responses are complex behaviors requiring processing by limbic and cortical circuits in the brain, the appearance of which indicates the subject’s motivation and arousal to make the aversive sensations cease, by licking the affected tissue (e.g. engaging spinal gate mechanisms), protecting the tissue, or seeking an escape route (Blanchard & Blanchard, 1969; Bolles, 1970; Bolles & Fanselow, 1980; Darwin, 1872; Estes, 1948; Estes & Skinner, 1941; Fanselow, 1982; Koch, 2004; Mogil, 2009; Rescola, 1968; Rescola & Lolordo, 1965; Skinner, 1938; Wiech & Tracey, 2013; Woolf, 1984). Most assuredly, the circuits encoding the affective dimension of pain perception versus those executing the subsequent affective-motivational behaviors are not performing the same computation or neural function. But it is likely these individual micro-circuits are connected into larger networks, where disruption of the affective encoding circuits would alter downstream motivational circuits and in turn behavioral actuators (e.g. spinal motor neurons and muscles). We do not yet have a neuroscientific set of definitions for these different operations within each circuit, but this is something to strive toward.

This approach at stringing together behavioral motifs, such as attending to the affected tissue by licking or protectively guarding, highlights the necessity of nociceptive processing in affective circuits to flexibly construct these reactive behaviors. For example, microendoscopic calcium imaging in the BLA revealed that nociceptive activity correlated with the onset of these affective-motivational behaviors. These BLA activities persisted until the animal ended its active-escape or attending behavioral routine, likely reflecting an action-outcome value updating process that maintains the engagement of motivational behaviors until nociception stops or is predicted to stop in the near future. In separate studies, we found that chemogenetic inhibition of BLA nociceptive neurons robustly and selectively reduced affective but not reflexive behaviors. Thus, using these behavioral measures of pain affective-motivational behavior we were able to directly link in-the-moment affective behaviors with underlying neural activities.

One of the current favored methodologies for purportedly evaluating the affective component of pain is a modified Conditioned Place Preference (CPP) paradigm, first developed by Sufka (Sufka, 1994)and refined by Porreca and colleagues (T. King et al., 2009). CPP is limited in its description of “spontaneous” pain affect because it conflates the associative learning process with the affective state being associated to the environment. Broadly, many brain regions show activity to both conditioned and unconditioned stimuli (Baeg et al., 2009; Hyman et al., 2017; Monosov, 2017), which makes it nearly impossible to tell on the post-conditioning test day which process (or both) was blocked by the administered analgesic. Moreover, using reward-seeking relief from chronic pain as an indirect readout of aversiveness of the resting state relies on the assumption that motivational-reward processes and cognitive flexibility remain intact. However, chronic pain causes dysfunction in reward-related processes, necessitating careful interpretation (Berryman et al., 2013; Moriarty et al., 2011, 2017; Schwartz et al., 2014). Thus, the CPP assay can assess the effectiveness of an intervention to motivate behavioral change, but this is fundamentally different from the affective state resulting in nociception. Instead, an automated brain imaging and behavioral platform leveraging affective-motivational pose-tracking could more objectively determine the animal’s in-the-moment behavioral responses to noxious events., By quantifying only unconditioned responses, one can begin to dissect the discrete networks processing cues and outcomes by linking the activity of nociceptive neurons to behavior, thereby separating the sensory-discriminative from affective-motivational dimensions of pain perception.

A few recent attempts to automate nociceptive behavior for withdrawal metrics (Abdus-Saboor et al., 2019; J. Jones et al., 2020) have pushed the field forward, but still rely on anthropomorphic, supervised labeling of video frames for network training. Indeed, rodents are not small humans, and humans remain poor rodent psychologists. Thus, unsupervised deep-learning of video-recorded behaviors derived from pose-tracking systems (e.g. DeepLabCut and Social LEAP Estimates Animal Poses (SLEAP)) can improve upon this behavioral classification system, allowing for the discovery of a modular structure of rodent pain-affective behavior (Mathis et al., 2018; Pereira et al., 2019). These generative statistical models, including Motion Sequencing (MoSeq) and Behavioral Segmentation of Open-field In DeepLabCut (B-SOiD), parse behavior into units of action based upon underlying statistical structure without relying upon explicit human labeling (Hsu & Yttri, 2019; Wiltschko et al., 2015). In line with ethological theory, pain affective behavior, like nearly all behavior, is constructed of repeated, stereotyped, and objectively quantifiable behavioral motifs.

Facial grimace is another recent measure of pain affect. Indeed, humans and animals partly communicate their inner mental life through facial expressions. Darwin asserted that the ability to communicate one’s emotional experience through facial expressions, including pain and pleasure, is a fundamental aspect of social world (Ekman, 2009). As a first step toward understanding this important question, Mogil and colleagues developed the original Mouse Grimace Scale that relied on human-scored single video-frames on a 0-2 scale (Langford et al., 2010). This process was later semi-automated by a supervised neural network trained on human-selected video frames from mice exposed to noxious stimuli (Tuttle et al., 2018). In a refined methodology from Gogolla and colleagues, the team used a machine-learning algorithm to classify facial features of not only pain, but disgust, pleasure, fear and malaise, and to then link those facial behaviors to noxious-responsive ensembles in the insular cortex via 2-photon calcium imagining in awake, behaving mice (Dolensek et al., 2020). This is an elegant demonstration of how combining machine deep-learning for facial-feature detection or bodily pose-estimation with single-neuron imaging and optogenetic manipulations can accelerate the decoding of neuronal circuit functions and mechanisms that generate internal pain affective experiences, and their outward behavioral responses.

1 Like