Abstract
Sensory stimuli can trigger an orienting reflex (response) by which animals move the head to position their sensors (e.g., eyes, pinna, whiskers). Orienting responses may be important to evaluate stimuli that call for action (e.g., approach, escape, ignore), but little is known about the dynamics of orienting responses in the context of goal-directed actions. Using mice of either sex, we found that, during a signaled avoidance action, the orienting response evoked by the conditioned stimulus (CS) consisted of a fast head movement containing rotational and translational components that varied substantially as a function of the behavioral and underlying brain states of the animal set by different task contingencies. Larger CS-evoked orienting responses were associated with high-intensity auditory stimuli, failures to produce the appropriate signaled action, and behavioral states resulting from uncertain or demanding situations and the animal's ability to cope with them. As a prototypical orienting neural circuit, we confirmed that the superior colliculus controls and codes the direction of spontaneous exploratory orienting movements. In addition, superior colliculus activity correlated with CS-evoked orienting responses, and either its optogenetic inhibition or excitation potentiated CS-evoked orienting responses, which are likely generated downstream in the medulla. CS-evoked orienting responses may be a useful probe to assess behavioral and related brain states, and state-dependent modulation of orienting responses may involve the superior colliculus.
SIGNIFICANCE STATEMENT Humans and other animals produce an orienting reflex (also known as orienting response) by which they rapidly orient their head and sensors to evaluate novel or salient stimuli. Spontaneous orienting movements also occur during exploration of the environment in the absence of explicit, salient stimuli. We monitored stimulus-evoked orienting responses in mice performing signaled avoidance behaviors and found that these responses reflect the behavioral state of the animal set by contextual demands and the animal's ability to cope with them. Various experiments involving the superior colliculus revealed a well-established role in spontaneous orienting but only an influencing effect over orienting responses. Stimulus-evoked orienting responses may be a useful probe of behavioral and related brain states.
Introduction
In alert animals, an unexpected, novel, or salient sensory stimulus leads to a well-known orienting response by which animals move their head and sensors in an apparent effort to evaluate the stimulus (Sokolov, 1963). Spontaneous orienting also occurs during active exploration of the environment in the absence of salient overt stimuli, but it is not clear whether stimulus-evoked and spontaneous orienting share the same neural substrate. Since Pavlov identified orienting responses, they have been intensely investigated in humans as a functional unconditioned response system that comprises numerous motor and physiological (e.g., vascular, respiratory, skin conductance, EEG) reactions (e.g., Ranganath and Rainer, 2003; Friedman et al., 2009). However, less is known about stimulus-evoked orienting responses in mice, an animal model amenable to studying the underlying neural circuits.
A decision to act usually follows stimulus-evoked orienting responses. For instance, during a learned sensory guided locomotor behavior in rodents, known as signaled active avoidance, the conditioned stimulus (CS) onset typically evokes a head orienting response that is dissociable from the ensuing avoidance action (e.g., Hormigo et al., 2021b; Zhou et al., 2022). However, little is known about CS-evoked orienting responses in the context of signaled avoidance behaviors. One possibility is that orienting responses adapt to changes in the meaning of the evoking CS, reflecting the significance of the stimulus. Alternatively, orienting responses may vary in relation to the demands of the situation and the animal's ability to cope with it, which affects the behavioral and associated brain state (neural circuits operating modes) (Castro-Alamancos and Connors, 1996; Castro-Alamancos, 2004a, 2010; Lee and Dan, 2012) of the animal during learned actions, including signaled avoidance (Castro-Alamancos, 2004b). Behavioral states can be conceptualized as a complex construct that includes overlapping variable dimensions (e.g., vigilance, emotion, motivation) that change with contextual demands (defined by levels of uncertainty, danger, control, predictability, resources, etc.) and the animal's coping abilities. In the present study, we evaluated how CS-evoked orienting responses adapt as a function of variations in CS meaning and/or behavioral states set by task contingencies.
The superior colliculus is intimately involved in signaled active avoidance (Cohen and Castro-Alamancos, 2007, 2010b,c; Hormigo et al., 2019), as well as other action selection and cognitive processes (e.g., Krauzlis et al., 2013; Yilmaz and Meister, 2013; Shang et al., 2015; Basso and May, 2017). Moreover, spontaneous and stimulus-evoked superior colliculus firing varies with the behavioral state of the animal (Cohen et al., 2008; Cohen and Castro-Alamancos, 2010b). The superior colliculus has been intensely studied as an essential component of the neural circuitry controlling orienting movements (Dean et al., 1989; Gandhi and Katnani, 2011; Corneil and Munoz, 2014), particularly eye movements (saccades) in head-fixed primates (Hikosaka and Wurtz, 1983a, 1983b; Sparks, 1986; Wurtz and Hikosaka, 1986; Hikosaka et al., 2000). Its layered organization, consisting of visual maps in superficial layers and multisensory maps in the intermediate layers, which are arranged relative to motor outputs in deeper layers (Schiller and Stryker, 1972; Stein et al., 1975; Sparks, 1986; Stein and Meredith, 1993), is well suited to support orienting, but the details remain to be established. While the activity of superior colliculus neurons has been studied in behaving rodents (e.g., Felsen and Mainen, 2008; Cohen and Castro-Alamancos, 2010b,c; L. Wang et al., 2022), only a few recent studies have related this neural activity to tracked orienting head movements. Thus, superior colliculus neurons discharge in association with contraversive head orienting (Wilson et al., 2018; Masullo et al., 2019); conversely, unilateral electrical or optogenetic excitation of these neurons drives contraversive head movements (Dean et al., 1986; Masullo et al., 2019; Isa et al., 2020). While the superior colliculus is under inhibitory control by GABAergic afferents, such as those from the SNr (Di Chiara et al., 1979), which can control head orienting (Hormigo et al., 2021a), little is known about superior colliculus inhibition in the context of CS-evoked or spontaneous exploratory orienting.
In mice of both sexes, we measured CS-evoked orienting during successive signaled avoidance procedures that change the meaning of the evoking stimulus and/or the behavioral state of the animal by adjusting task contingencies. We also used a combination of optogenetics, electrophysiology, and fiber photometry to manipulate and monitor the activity of superior colliculus neurons during CS-evoked orienting responses in the context of signaled avoidance and during spontaneous exploratory orienting behaviors.
Materials and Methods
Experimental design and statistical analysis
All procedures were reviewed and approved by the institutional animal care and use committee and conducted in adult (>8 weeks) male and female mice. The results from both sexes were combined since there is no sex difference measured in basic active avoidance behavior (Hormigo et al., 2019).
Experiments involved a repeated-measures design in which the mice or cells serve as their own controls (comparisons within groups), but we also compared experimental mice to other groups (e.g., to No Opsin controls) and different groups of experimental animals to each other as noted (comparisons between groups). For comparisons within groups, we tested for a main effect using a repeated-measures ANOVA or a linear mixed-effects model in which procedures (behavioral task), CS, or optogenetic stimulus (with as many levels as conditions tested; e.g., blue light patterns) is a fixed-effect and Sessions are a random-effect nested within the Subjects (Data ∼ Stimulus + (1|Subjects/Sessions); as per lme4 syntax in R) followed by comparisons with Tukey's test. For comparisons between different groups, we used the same approach but included the Group as an additional fixed-effect (Data ∼ Group × Stimulus + (1|Subjects/Sessions)). Using the SEs derived from the model, Tukey tests were conducted for the effect of the fixed-effect factors (within group comparisons) or for the Group-Stimulus interaction (between group comparisons). We report the Tukey values for the relevant multiple comparisons. In addition, in many cases, we performed the same comparisons (both within groups and between groups) using a bootstrap approach by randomly sampling with replacement (1000-10,000 times) from all the values (regardless of the conditions or groups) and determining the probability that the difference (or a larger difference) between the conditions or groups occurs by chance (these p values are reported inside brackets, []). For the optogenetics experiments in an open field, we conducted a power analysis with OriginLab Pro using the measured means difference variability. This revealed that 3 animals in which we conducted ∼5 identical daily sessions per animal (15 sessions in 3 mice) was sufficient to detect a significant change (∼10° change in head bias) with a power of 0.99 (p = 0.05). This was used as the bare minimum number of animals and sessions in the optogenetics open field experiments.
To enable rigorous approaches, we maintain a centralized metadata system that logs all details about the experiments and is engaged for data analyses (Castro-Alamancos, 2022). Moreover, during daily behavioral sessions, computers run experiments automatically using preset parameters logged for reference during analysis. Analyses are performed using scripts that automate all aspects of data analysis from access to logged metadata and data files to population statistics and graph generation.
Strains and adeno-associated viruses (AAVs)
The following AAVs (injected undiluted) and mouse strains were used to manipulate the activity of neurons with optogenetics by expressing ChR2 or Arch, or to measure the activity of neurons with fiber photometry by expressing GCaMP6f/7f. Optical fibers were implanted in the superior colliculus, unless otherwise indicated. As No-Opsin controls, we injected AAV8-hSyn-EGFP (Addgene, titers: 4.3 × 1012 GC/ml by qPCR) or nil in the superior colliculus. All the optogenetic methods used in the present study have been validated in previous studies using slice and/or in vivo electrophysiology (Hormigo et al., 2016, 2019, 2021a,c).
To inhibit the superior colliculus, we used three strategies: (1) In Vgat-SNr-ChR2 mice, we excite GABAergic afferent fibers in the superior colliculus. To express ChR2 in these GABAergic afferents, we injected AAV5-EF1a-DIO-hChR2(H134R)-eYFP (UPenn Vector Core or Addgene, titers: 1.8 × 1013 GC/ml by qPCR) in the SNr of Vgat-cre mice (Jax 028862; B6J.129S6(FVB)-Slc32a1tm2(cre)Lowl/MwarJ) mice. (2) In Vgat-SC-ChR2 and Vgat-ChR2 mice, we excite GABAergic neurons in the superior colliculus, which inhibits glutamatergic superior colliculus neurons. To express ChR2 in GABAergic neurons, we injected AAV5-EF1a-DIO-hChR2(H134R)-eYFP (UPenn Vector Core or Addgene, titers: 1.8 × 1013 GC/ml by qPCR) in the superior colliculus of Vgat-cre mice (Vgat-SC-ChR2), or crossed Vgat-cre and Ai32 (Jax 024109; B6.Cg-Gt(ROSA)26Sortm32(CAG-COP4*H134R/EYFP)Hze/J) mice (Vgat-ChR2). (3) In CaMKII-SC-Arch and Vglut-Arch mice, we directly inhibit glutamatergic neurons in the superior colliculus. To express Arch in glutamatergic superior colliculus neurons, we crossed Vglut-cre (Jax 016963; Slc17a6tm2(cre)Lowl/J) and Ai40D (Jax 021188; B6.Cg-Gt(ROSA)26Sortm40.1(CAG-aop3/EGFP)Hze/J) mice (Vglut-Arch), or we targeted CaMKII-expressing neurons, which are primarily glutamatergic (X. Wang et al., 2013), by injecting AAV5-CaMKIIa-eArchT3.0-EYFP (UNC Vector Core, titers: 4 × 1012 GC/ml by Dot Blot) into the superior colliculus of C57BL/6J mice (Jax 00664) (CaMKII-SC-Arch).
To excite the superior colliculus, we used three strategies: (1) In CaMKII-SC-ChR2 mice, we directly excite glutamatergic neurons in the superior colliculus. To express ChR2 in glutamatergic neurons, we injected AAV5-CaMKIIa-hChR2(H134R)-EYFP (UNC Vector Core, titers: 6.24 × 1012 GC/ml by Dot Blot) into the superior colliculus of C57BL/6J mice. (2) In Chat-ChR2 mice, we excite cholinergic cells and fibers in the superior colliculus, which has local excitatory effects (Castro-Alamancos and Favero, 2016). To express ChR2 in cholinergic cells, we crossed Chat-cre (Jax 006410; B6;129S6-Chattm2(cre)Lowl/J) and Ai32 mice. (3) In Vgat-Arch mice, we inhibit superior colliculus GABAergic cells, which disinhibits superior colliculus glutamatergic cells. To express Arch in GABAergic neurons, we crossed Vgat-cre and Ai40D mice.
To measure calcium signals in superior colliculus neurons, we injected C57BL/6J mice (Jax 00664) with AAV5-CaMKII-GCaMP6f-WPRE-SV40 (UPenn Vector core, titers: 1.53 × 1013 GC/ml).
Surgeries
Optogenetics and fiber photometry experiments involved injecting 0.2-0.4 µl AAVs per site during isoflurane anesthesia (∼1%). Animals received carprofen after surgery. The stereotaxic coordinates for injection sites (in mm from bregma; lateral from the midline; ventral from the bregma-λ plane) are: SNr (3.3 posterior; 1.4-1.5; 4-4.1), and superior colliculus (4 posterior; 1; 1.75). The coordinate ranges reflect different animals that were grouped together because the slight coordinate differences produced similar effects.
In optogenetics and fiber photometry experiments, a single or dual optical fiber (200 µm in diameter) was implanted unilaterally (photometry) or bilaterally during isoflurane anesthesia in the superior colliculus (4 posterior; 1-1.5; 1.5-1.8), and held in place with a combination of screws, cyanoacrylate, and dental cement.
In unit recording experiments, a bundle of a several electrodes including a ground, analog reference, and a few tungsten sharp (edged) electrodes (1-2 mΩ) was implanted in the superior colliculus using the same coordinates as for the optical fibers. The ground and analog reference were placed outside the superior colliculus, and a blunt tungsten electrode served as a digital reference located near the recording electrodes.
Optogenetics
The implanted optical fibers were connected to patch cables using sleeves. A black aluminum cap covered the head implant and completely blocked any light exiting at the ferrule's junction. Furthermore, the experiments occurred in a brightly lit cage that made it difficult to detect any light escaping the implant. The other end of the patch cables was connected to a dual light swivel (Doric lenses) that was coupled to a green laser (520 nm; 100 mW) to activate Arch or a blue laser (450 nm; 80 mW) to activate ChR2. Unless otherwise noted, the behavioral experiments used green light between 25 and 35 mW and blue light between 1 and 3 mW. Power is regularly measured by flashing the connecting patch cords onto a light sensor, with the sleeve on the ferrule.
Fiber photometry
We used a 2-channel (465 and 405 nm) fiber photometry system (Doric Lenses) with alternating (20-60 and 20-50 µW, respectively) pulses of light excitation at 100 Hz (per each 10 ms, 465 is on for 3 ms, and 2 ms later 405 is on for 3 ms). The emission peak signals (525 and 430 nm for GCaMP6f and control emissions) evoked by the 465 nm and 405 pulses were acquired at 5-20 kHz and measured at the end of each pulse. To calculate Fo, the 430 signal was scaled to the 525 signal (F) using the slope of the linear fit. Finally, F/Fo was calculated with the following formula: (F – Fo)/Fo and converted to Z scores. Because of the nature of the behavior studied, a swivel is essential. We used a rotatory-assisted photometry system that has no light path interruptions (Doric Lenses). In addition, an aluminum cap was implanted during surgery and black acrylic powder was mixed in the dental cement to assure that ambient light was not leaking into the implant and reaching the optical fiber; this was tested in each animal by comparing fluorescence signals in the dark versus normal cage illumination.
Unit recordings
Mice were connected to a rotatory-assisted electrical swivel by an ultra-light cable and digital headstage (<1 g) and were free to move in an open field. Activity from all the electrodes was continuously recorded as wide-band, high-pass, and low-pass filtered signals. Unit activity was threshold detected and cluster cut using Offline Sorter (Plexon) or our automated scripts. As our criteria for single units is based on high impedance, movable electrodes, which yield perfectly isolated spikes, we designated most of our recordings as unit clusters; only a group of very well-isolated clusters with very high signal-to-noise ratios were denoted, single units. The optrode recordings were obtained with similar methods, but the electrodes were attached to an optogenetic fiber and protruded (100-300 µm) below it.
Orienting and video tracking
During open field experiments, mice are placed in a circular open field (10 inch diameter) that was illuminated from the bottom. All mice in the study (open field or shuttle box) were continuously video tracked (30-100 FPS) in synchrony with the procedures and other measures. We automatically tracked head movements with two color markers attached to the head connector: one located over the nose and the other between the ears. The coordinates from these markers form a line (head midline) that serves to derive several instantaneous movement measures per frame. Overall head movement was separated into rotational and translational components. Rotational movement was the angle formed by the head midline between succeeding video frames multiplied by the radius. Translational movement resulted from the sum of linear (forward vs backward) and sideways movements. Linear movement was the distance moved by the ears marker between succeeding frames multiplied by the cosine of the angle formed by the line between these succeeding ear points and the head midline. Sideways movement was calculated as linear movement, but the sine was used instead of the cosine. We also estimated head tilt by measuring the distance between the head markers per frame, which varies as mice tilt the head. Pixel measures were converted to metric units using calibration. Thus, a total of six absolute head movements measures expressed as speed (cm/s) were derived, including overall, rotational, translational, linear, sideways, and tilt. In AA tasks, we used absolute measures, unless direction was considered. Relative rotational and sideways movements flatten when averaged over trials because mice can turn either left or right to accomplish the task successfully, and the CS is delivered binaurally. Also, relative linear movement was mostly forward as backwards movement is minimal in the task. We then used the time series to extract window measurements around events (e.g., CS presentations). The baseline window was the average of a 0.5-1 s period before an event. For the response windows (e.g., orienting, avoidance), we calculated the maximum (Max), minimum (Min), and/or extrema (Peak) of the instantaneous speeds per a defined period after the event. Moreover, the Max, Min, and Peak were corrected by subtracting the baseline speed. Measurements were obtained from single-trial traces and/or from traces averaged over a session. In addition, we obtained the direction of the rotational movement with a Head Angle or bias measure, which was the accumulated change in angle of the head per frame (vs the previous frame) zeroed by the frame preceding the stimulus onset or event (this is equivalent to the rotational speed movement in degrees). The time to peak is when the extrema occur versus event onset.
To detect spontaneous orienting from the head tracking, we applied a local maximum algorithm to the continuous head angle measure. Every point is checked to determine whether it is the maximum or minimum among the points in a range of 0.5 s before and after the point. If the change in angle of this point is >10 degrees, it is considered a turn in the direction of the sign. We further sorted turns based on the timing of previous turns.
Active avoidance task in a shuttle box
Mice were trained in a signaled active avoidance task, as previously described (Hormigo et al., 2016, 2019). During an active avoidance session, mice are placed in a standard shuttle box (16.1 inch × 6.5 inch) that has two compartments separated by a partition with side walls forming a doorway that the animal has to traverse to shuttle between compartments (see Fig. 1A). A speaker is placed on one side, but the sound fills the whole box, and there is no difference in behavioral performance (signal detection and response) between sides. A trial consists of a 7 s avoidance interval followed by a 10 s escape interval. During the avoidance interval, an auditory CS (8 kHz, 85 dB) is presented for the duration of the interval or until the animal produces a conditioned response (avoidance response) by moving to the adjacent compartment, whichever occurs first. If the animal avoids by moving to the next compartment, the CS ends, the escape interval is not presented, and the trial terminates. However, if the animal does not avoid, the escape interval ensues by presenting white noise and a mild scrambled electric footshock (0.3 mA) delivered through the grid floor of the occupied half of the shuttle box. This unconditioned stimulus (US) readily drives the animal to move to the adjacent compartment (escape response), at which point the US terminates, and the escape interval and the trial ends. Thus, an avoidance response will eliminate the imminent presentation of a harmful stimulus. An escape response is driven by presentation of the harmful stimulus to eliminate the harm it causes. Successful avoidance warrants the absence of harm. Each trial is followed by an intertrial interval (duration is randomly distributed; 25-45 s range), during which the animal awaits the next trial. We used three variations of the basic signaled active avoidance procedure termed AA1, AA2, and AA3 (see Fig. 1B).
In AA1, mice are free to cross between compartments during the intertrial interval; there is no consequence for intertrial crossings (ITCs).
In AA2, mice receive a 0.2 s footshock (0.3 mA) and white noise for each ITC. Therefore, in AA2, mice must passively avoid during the intertrial interval by inhibiting their tendency to shuttle between trials. Thus, during AA2, mice perform both signaled active avoidance during the signaled avoidance interval (like in AA1) and unsignaled passive avoidance during the unsignaled intertrial interval.
In AA3, mice are subjected to a CS discrimination procedure in which they must respond differently to a CS1 (8 kHz tone at 85 dB) and a CS2 (4 kHz tone at 70 dB) presented randomly (half of the trials are CS1). Mice perform the basic signaled active avoidance to CS1 (as in AA1 and AA2), but also perform signaled passive avoidance to CS2, and ITCs are not punished. In AA3, if mice shuttle during the CS2 avoidance interval (7 s), they receive a 0.5 s footshock (0.3 mA) with white noise and the trial ends. If animals do not shuttle during the CS2 avoidance interval, the CS2 trial terminates at the end of the avoidance interval (i.e., successful signaled passive avoidance).
There are three main variables representing task performance. The percentage of active avoidance responses (% avoids) represents the trials in which the animal actively avoided the US in response to the CS. The response latency (latency) represents the time (seconds) at which the animal enters the safe compartment after the CS onset; avoidance latency is the response latency only for successful active avoidance trials (excluding escape trials). The number of crossings during the intertrial interval (ITCs) represents random shuttling because of locomotor activity in the AA1 and AA3 procedures, or failures to passively avoid in the AA2 procedure. The sound pressure level (SPL) of the auditory CS were measured using a microphone (PCB Piezotronics 377C01) and amplifier (×100) connected to a custom LabVIEW application that samples the stimulus within the shuttle cage as the microphone rotates driven by an actuator controlled by the application.
Histology
Mice were deeply anesthetized with an overdose of ketamine. Upon losing all responsiveness to a strong tail pinch, the animal was decapitated and the brain was rapidly extracted and placed in fixative or used for slice recordings before placement in fixative. Recordings in slices were performed to validate the optogenetic effects, as reported in previous studies (Hormigo et al., 2016, 2019, 2021a,c). The fixed brains were sectioned (100 µm sections) in the coronal or sagittal planes. Sections were mounted on slides, cover-slipped with DAPI mounting media, and photographed using a fluorescent microscope.
Results
CS-evoked orienting responses
Using head markers, we measured head movements in any direction evoked by a binaural auditory tone CS in the context of learned signaled active avoidance behavior in a shuttle box (Fig. 1A), wherein mice learn to avoid an aversive US by shuttling between two compartments in a cage during an avoidance interval (7 s) signaled by the CS. If the animal fails to avoid during the CS, the US is presented causing a rapid escape to the other compartment, which eliminates the US and initiates a random intertrial interval until the next trial. As described ahead, we used different avoidance procedures that vary in difficulty or demand (Fig. 1B, AA1-AA3). In the basic signaled active avoidance procedure termed AA1, mice are free to shuttle between the cage compartments during the random intertrial interval that separates CS presentations (ITCs; gray filled bars; Fig. 1B) and have no consequence during AA1 (Hormigo et al., 2021b; Zhou et al., 2022). During all procedures, we derived six movement measures, the overall head movement (Fig. 1C), which consisted of rotational and translational components (Fig. 1D), where the translational movement is further subdivided into linear and sideways components, and we also estimated head tilt from the 2D tracking. All movement measures are plotted as speed (cm/s) and angles are plotted in degrees. For each movement measure, we calculated three basic epochs around the CS presentations. The baseline movement was the average head speed measured 500 ms before CS onset. The orienting response movement was the maximum (Max) head speed measured during the first 500 ms after CS onset corrected by the baseline movement. In some cases, noted below, we also show the minimum (Min) for this epoch, which becomes relevant when a CS drives a significant inhibitory (arresting) orienting movement. The avoidance movement was the average head speed measured between 500 ms after CS onset and the end of a successful avoidance trial corrected by the baseline movement.
CS-evoked orienting responses in the context of signaled active avoidance. A, Arrangement of the shuttle box used during signaled avoidance tasks. B, Schematic represents the timing of avoidance, escape, and intertrial intervals during the three avoidance procedures (AA1, AA2, and AA3). AA1 and AA2 are signaled avoidance tasks, where ITCs are punished in AA2. AA3 is a discrimination procedure. CS2 (4 kHz) in AA3 signals to passively avoid (“do not cross or get US, if you cross”), while CS1 (8 kHz) signals to actively avoid (“cross or get US when the escape interval begins, if you don't cross”). C, CS-evoked orienting response measured by tracking overall head speed. The orienting response is triggered by the onset of the auditory stimulus that signals the avoidance interval. The orienting response movement is followed by the avoidance movement (i.e., when the animal moves to avoid the US). The traces show the first training session (black) and the fifth session (red) for naive mice trained in AA1. Right, Close-up of the left panel. D, Rotational and translational components of the overall head speed shown in C. E, Linear (forward) and sideways components of the translational movements shown in D. Overall head tilt is also shown.
We trained naive mice (n = 10) in AA1 during five sessions. Figure 1C shows average traces of overall head movement for the first (black) and fifth (red) AA1 session. On the first session, there is a prominent movement of the head driven by the CS during the orienting response window; perusal of the video frames revealed clear head movements to either side resembling a typical orienting response (“what is it?”) movement. This was followed by the avoidance response movement, which occurs later. However, on the fifth session, the orienting movement was mostly absent while the avoidance movement became more prominent as the animals learned to avoid. Figure 1D,E shows the corresponding movement components, all of which adapted significantly between the sessions. Figure 2 shows measurements of overall, rotational, and translational movement for the three response windows (Fig. 2B) together with the corresponding behavioral performance per session (Fig. 2A), which is represented by the percentage of active avoids (black circles; Fig. 2A), avoidance latencies (orange triangles; Fig. 2A), and the number of ITCs during the intertrial interval (gray bars). As is typical during AA1 learning, the number of avoidance responses (Tukey t(36) =16.2, p < 0.0001, first vs fifth session), and ITCs (Tukey t(36) = 4.9, p = 0.01, first vs fifth session) increased over the sessions (Fig. 2A).
CS-evoked orienting responses habituate as naive mice learn the AA1 signaled active avoidance procedure. A, Behavioral performance during five AA1 sessions showing the percentage of active avoids (black circles), avoidance latency (orange triangles), and number of ITCs (gray bars). B, Baseline (−0.5 to 0 s window vs CS onset), orienting response (0-0.5 s window), and avoidance response (0.5-5 s) measured as overall, rotational, and translational movement speed (cm/s in all figures). Orienting and avoidance movements are corrected by the baseline movement. Orienting responses were large on the first training session but habituated rapidly as mice acquired signaled avoidance. Population data and traces are mean ± SEM in all figures.
The movement measures changed as mice acquired the ability to avoid the US. Orienting responses were prominent during the first training session but habituated substantially over the following sessions as mice learned the task. Thus, overall (Tukey t(36) = 8.75, p < 0.0001, first vs fifth session or any other session), rotational (Tukey t(36) = 4.93, p = 0.01), and translational (Tukey t(36) = 4.72, p = 0.01) movements, but not tilt (Tukey t(36) = 2.63, p = 0.35) habituated. The translational movement habituation involved primarily the linear (Tukey t(36) = 4.18, p = 0.04), not the sideways (Tukey t(36) = 3.69, p = 0.08) component. Orienting response habituation was associated with an increase in overall baseline movement (Tukey t(36) = 8.37, p < 0.0001, first vs fifth session), reflected by every component, including rotational (Tukey t(36) = 8.3, p < 0.0001), translational (Tukey t(36) = 8.76, p < 0.0001), linear (Tukey t(36) = 8.31, p < 0.0001), sideways (Tukey t(36) = 9, p < 0.0001), and tilt (Tukey t(36) = 8.97, p < 0.0001) movements.
Overall avoidance movement tended to increase over sessions, but this was not significant (Tukey t(36) = 1.75, p = 0.72, first vs fifth session), indicating that from the first session mice moved more during the CS presentation but initially fail to move in the correct direction to avoid. This was made apparent when the movement components were dissociated, which revealed that only linear avoidance movement showed a tendency to increase with learning (Tukey t(36) = 4.2, p = 0.03), highlighting the forward motion of avoidance responses toward the adjacent (safe) compartment. In conclusion, as mice learn to actively avoid a harmful US, the initially large CS-evoked orienting responses habituate substantially; well-trained mice show decreased CS-evoked orienting responses compared with the initial training session.
CS-evoked orienting responses reflect task contingencies
The previous results indicate that stimulus-evoked orienting responses habituate in the context of learning, but little is known about how behavioral task contingencies affect orienting responses. During signaled active avoidance training, the stimulus that was initially neutral becomes a CS that predicts the harmful US (i.e., acquires new meaning). While learning this new contingency predicted by the CS, animals substantially adjust their behavioral state, which is evident by variations in external (movement and posture) and internal (neural circuit) measures, such as the level of alertness and related thalamocortical activation (Castro-Alamancos, 2004a,b); alertness and thalamocortical activation (which blunts rapid sensory adaptation) are maximal during AA1 learning but ease significantly as the coping behavior is acquired. Therefore, variations in orienting responses during signaled avoidance may reflect the change in significance of the evoking stimulus, as it transitions from neutral to conditioned, and/or may reflect behavioral state adjustments to cope with the situation. To test these possibilities, we measured CS-evoked orienting responses while mice performed three signaled avoidance procedures (AA1-AA3; Fig. 1B) that adjust stimulus significance and/or behavioral states.
Using naive mice (n = 15; Fig. 3), we first determined the native orienting response evoked by a meaningless (neutral) auditory stimulus presented without consequence (noUS; five sessions). This was followed by training with this same stimulus in AA1, AA2, and AA3 (Hormigo et al., 2021b; Zhou et al., 2022) (seven sessions per procedure). In AA1, the neutral auditory stimulus acquires new meaning by becoming a CS that predicts the US, which can be avoided by shuttling during the CS presentation. AA2 is identical to AA1 (CS significance remains constant), but ITCs are punished, resulting in a more cautious behavioral state as mice must not shuttle between CS presentations (unsignaled passive avoidance) (Zhou et al., 2022). AA3 is a more challenging discrimination procedure with two different CS presented randomly, and ITCs are not punished. CS1 is identical to the CS in AA1/AA2 (signaled active avoidance), but during CS2 mice must passively avoid punishment by not shuttling (signaled passive avoidance). Thus, AA3 presents a different situation where CS1 significance remains constant, but CS2 predicts an opposite contingency associated with the same action (shuttling), and ITCs have a different consequence.
CS-evoked orienting responses vary in association with the behavioral states set by No US, AA1, AA2, and AA3 procedures. A, Behavioral performance during the different procedures (panel details are as per Fig. 1). B, Baseline and orienting response (overall) movement, including the rotational and translational components of the orienting response during the different procedures. C, Overall head movement traces for different procedures. D, The rotational and translational components of the traces shown in C.
Figure 3 shows behavioral performance (Fig. 3A) and CS-evoked orienting (Fig. 3B,C) during the successive procedures. The procedures had a significant impact on overall baseline movement (one-way repeated-measures ANOVA; F(3,42) = 37, p < 0.0001), orienting movement (F(3,42) = 10.75, p < 0.0001), avoidance movement (F(3,42) = 56.23, p < 0.0001), percentage of active avoids (F(3,42) = 364.12, p < 0.0001), avoidance latencies (F(3,42) = 61.33, p < 0.0001), and number of ITCs (F(3,42) = 75.5, p < 0.0001). Focusing on the orienting movement, the procedures also significantly affected rotational (F(3,42) = 11.2, p < 0.0001), translational (F(3,42) = 10.45, p < 0.0001), linear (F(3,42) = 11.17, p < 0.0001), and sideways (F(3,42) = 6.99, p < 0.0001) movements, but not tilt (F(3,42) = 2.63, p = 0.06). Next, we evaluated these parameters over the sessions as mice transitioned between different procedures.
During No US sessions, mice ignore the neutral stimulus presentation since it has no consequence. Although shuttling during the stimulus presentation (7 s) turns it off, mice do not learn to do this, indicating that the auditory tone is not aversive. The active avoids and ITCs during No US sessions (Fig. 3A) reflect shuttling associated with exploratory movement.
During the AA1 procedure, the US occurs at the end of the stimulus presentation if animals do not shuttle. Mice rapidly learn to avoid the US by shuttling, but the contingency change obviously generates significant uncertainty about what is occurring. Compared with the last No US sessions, during the first AA1 session, there was a sharp increase in the overall CS-evoked orienting response (Tukey t(14) = 5.06, p = 0.003) involving mainly rotational (Tukey t(14) = 5.56, p = 0.001), but also some translational (Tukey t(14) = 3.5, p = 0.02) movement. This was associated with an increase in active avoids (Tukey t(14) = 9.02, p < 0.0001) and a drop in ITCs (Tukey t(14) = 9.59, p < 0.0001) and overall baseline movement (Tukey t(14) = 11.2, p < 0.0001). By the last AA1 session, the overall orienting response (Tukey t(14) = 1.65, p = 0.26), ITCs (Tukey t(14) = 0.97, p = 0.34), and overall baseline movement (Tukey t(14) =1.89, p = 0.2) recover substantially and no longer differ compared with the last No US session while active avoids were maximized (Tukey t(14) = 23.2, p < 0.0001). Thus, the orienting response increase that occurred during the first AA1 session habituated over the following AA1 sessions; by the last AA1 session, the overall orienting response was reduced compared with the first AA1 session (Tukey t(14) = 3.84, p = 0.01). The orienting response habituation occurred concomitant with an increase in ITCs (Tukey t(14) = 8.58, p < 0.0001), overall baseline movement (Tukey t(14) = 5.8, p = 0.001), and active avoids (Tukey t(14) = 14.39, p < 0.0001). Thus, AA1 causes an initial large increase in the CS-evoked orienting response, which habituates as mice learn active avoidance.
There is significant change in behavioral state between the first AA1 and last AA1 sessions. The first session can be characterized as uncertain (things are changing unexpectedly) and stressful (harmful footshocks are occurring), which is associated with vigilant, anxious, and fearful states, generally considered negative. In contrast, while the environment and contingencies remain the same, by the last AA1 session, the mice have acquired coping skills to predict and control the US. Consequently, mice adjust their behavioral state by engaging in typical exploratory and resting behaviors that occur during predictable and stressless situations. The CS-evoked orienting response adjustments during AA1 may reflect changes in behavioral state and/or in CS meaning. To distinguish between these options, we trained mice in AA2 and AA3.
During AA2, the CS meaning remains constant, but the unsignaled passive avoidance contingency during the intertrial interval produces a state of increased caution, which is evident in delayed avoidance latencies (Zhou et al., 2022). Compared with the last AA1 sessions, during the third AA2 session, when avoidance latency becomes clearly delayed (Tukey t(14) = 8.09, p < 0.0001), there was an increase in the overall CS-evoked orienting responses (Tukey t(14) = 4.62, p = 0.005), which involved rotational (Tukey t(14) = 3.23, p = 0.03) and translational (Tukey t(14) = 4.74, p = 0.004) movements. This was associated with a substantial drop in ITCs (Tukey t(14) = 12.26, p < 0.0001) and overall baseline movement (Tukey t(14) = 7.01, p < 0.0001). These changes in orienting response, avoidance latency, and ITCs became persistent as they did not change between the third and last AA2 sessions. However, overall baseline movement recovered substantially during AA2 (Tukey t(14) = 3.21, p = 0.03, third vs last AA2 session), indicating that animals increase their exploratory movement within the compartment they are located in (as they do not produce ITCs). Therefore, the changes in CS-evoked orienting response are not related to the baseline movement changes.
During AA3, mice continue to actively avoid in response to the same CS (CS1) but must learn to passively avoid during a different CS2. This new discrimination challenge immediately increased the overall CS1-evoked orienting response compared with the last AA2 sessions (Tukey t(14) = 4.22, p = 0.009), which involved mostly rotational (Tukey t(14) = 4.45, p = 0.007), not translational (Tukey t(14) = 2.06, p = 0.17) movements. This did not affect overall baseline movement (Tukey t(14) = 2.68, p = 0.07) or ITCs (Tukey t(14) = 0.79, p = 0.58), supporting the idea that changes in CS-evoked orienting response are not simply reflecting changes in baseline movement. Over the course of AA3 training, the number of ITCs (Tukey t(14) = 5.61, p = 0.001, first vs last AA3 session) and overall baseline movement recovered substantially (Tukey t(14) = 7.51, p < 0.0001), but the overall CS1-evoked orienting response remained persistently elevated (Tukey t(14) = 1.34, p = 0.35). Intriguingly, the CS2-evoked orienting response was inhibitory (i.e., head speed was reduced). This could be related to the CS2 distinct meaning and/or its lower intensity (SPL), which we addressed next. To compute this inhibitory orienting response feature, we measured the Min of the CS2-evoked responses.
Relation of CS-evoked orienting with SPL and the ensuing action
To determine whether the amplitude of the CS-evoked orienting response was a function of the auditory tone SPL in the context of signaled avoidance, we continued training the mice in AA3, while also randomly presenting six short pips (100 ms) at different SPLs (70, 77, and 85 dB) and at the CS1 and CS2 tone frequencies without consequence. Figure 4A,C shows the orienting responses evoked by the eight stimuli, including CS1 and CS2 proper. The earliest component of the overall CS-evoked orienting response (Max; Fig. 4A,C, pips) increased with tone intensity for both CS1-related (8 kHz; F(2,28) = 14.38, p < 0.0001) and CS2-related (4 kHz; F(2,28) = 9.17, p < 0.0001) pips. The CS1-related pips increased both the rotational (8 kHz; F(2,28) = 5.19, p = 0.01) and translational (8 kHz; F(2,28) = 19.2, p ≤ 0.0001) movements, while the CS2-related pips increased rotational (4 kHz; F(2,28) = 4.45, p = 0.02) but not translational (4 kHz; F(2,28) = 0.44, p = 0.64) movements. However, there was no difference between the CS1-related and CS2-related pips (Tukey t(14) = 1.1, p = 0.44). In contrast, the inhibitory component (Min; Fig. 4A,C, pips) remained mostly invariant as a function of SPL for CS1-related (F(2,28) = 0.55, p = 0.57) and CS2-related (F(2,28) = 1.1, p = 0.34) pips, but was stronger for CS2-related than CS1-related pips (Tukey t(14) = 4.6, p = 0.005). Thus, CS2 (passive avoidance contingency) evokes a stronger inhibitory orienting response than CS1 (active avoidance contingency), indicating that the orienting response prepares the animal for the ensuing action. Both SPL and the ensuing action affect CS-evoked orienting responses.
CS-evoked orienting responses depend on SPL and prepare for the ensuing action. A, Head speed traces evoked by the CS1 (8 kHz) and CS2 (4 kHz) pips, including the CS1 and CS2 proper during performance of AA3. Auditory pips were evoked at three different SPLs for the CS1 and CS2 frequencies. B, Head speed traces of CS1 and CS2 trials in which mice performed the correct response (active avoids for CS1 and passive avoids for CS2) or the incorrect response (escapes for CS1 and errors for CS2). C, Population measures of the orienting responses (Max and Min) evoked by the pips shown in A decomposed into rotational and translational movement components. *p < 0.05, significant differences between pips at the same frequency or between frequencies. Rightward panels represent the same measures for avoidance trials classified as correct versus incorrect responses during AA1/AA2 or CS1 in AA3 (Active avoids vs Escapes), and CS2 in AA3 (Passive avoids vs Errors). These populations measures compare the CS-evoked orienting responses shown in B. *p < 0.05, significant differences between the correct and incorrect responses. #The differences for all trials (unclassified) between CS1 and CS2.
Relation of CS-evoked orienting with trial success
Finally, we determined whether the CS-evoked orienting response varied as a function of the trial success. For AA1, AA2, and AA3 (CS1), we compared active avoids to escapes (when mice fail to avoid). For AA3 (CS2), we compared passive avoids to errors (when mice shuttle during CS2). It is important to consider that, because of the high level of behavioral performance, the number of escapes and errors is small compared with successful responses, which may impact this analysis (e.g., higher variability of the failed response measures). Figure 4B,C shows that, in AA1, AA2, and AA3 (CS1) procedures, escapes were associated with larger overall CS-evoked orienting responses than active avoids (Tukey t(14) = 7.85, p < 0.0001), which involves both rotational (Tukey t(14) = 7.85, p < 0.0001) and translational (Tukey t(14) = 10.07, p < 0.0001) movements. The inhibitory CS2-evoked orienting response (Min) in AA3 evoked by CS2 was larger for errors than for correct passive avoids (Tukey t(12) = 4.35, p = 0.009), which involves mainly rotational (Tukey t(12) = 4.5, p = 0.007) but also translational (Tukey t(12) = 3.5, p = 0.02) movements (the Max measure followed the same pattern as the Min; Fig. 4C). Thus, failed trials in both signaled active and passive avoidance are associated with larger orienting responses. This indicates that orienting responses may reflect a negative behavioral and internal state that is not conducive to effective behavioral performance, such as inattentiveness, anxiety, or fear.
In conclusion, CS-evoked orienting responses consist of a fast head movement that is sensitive to SPL but may also include an inhibitory component by which the animal suppresses its ongoing head movement, which is less sensitive to SPL and become prominent when the CS signals the need to postpone/inhibit the action. CS-evoked orienting responses reflect changes in behavioral state associated with distinct task contingencies. Larger CS-evoked orienting responses are associated with negative behavioral states and poor behavioral performance typical of uncertain or demanding situations.
Where do mice orient to during signaled avoidance?
Next, we evaluated the direction of the CS-evoked orienting response (n = 25 mice) to determine where mice orient to in the context of signaled active avoidance (AA1, AA2, and AA3 CS1). We considered that mice might either orient toward the location of the CS source (speaker), the relevant location to accomplish the task, which in signaled active avoidance is the avoid/escape route (door), or simply move in a random or undetermined direction. We measured the change in distance between trial onset and orienting response offset in trials when the animal is on side of the shuttle box where the door and speaker are on opposite ends. This revealed whether mice closed the distance with the door or the speaker. Figure 5 shows that, during the orienting response, mice increased the distance to the speaker (one-sample test t(11,341) = 9.6, p < 0.0001) and accordingly decreased the distance to the door (one-sample test t(11,341) = 6.1, p < 0.0001). These findings make sense because the location of the speaker has no significance in signaled active avoidance. In contrast, the location of the avoid/escape route is highly significant. Moreover, when we separated trials by avoids and escapes, these effects were highly significant for avoids (p < 0.0001) but were absent for escapes (p > 0.6), although as shown in the previous section escapes tend to produce larger orienting responses. These results support the contention that CS-evoked orienting responses may have a role in preparing for the upcoming action signaled by the CS. Furthermore, we expect the direction of the orienting response to be different in different contexts, such as when the sound comes from a predator or when it signals a feeding location.
During signaled active avoidance, the CS-evoked orienting response orients the head toward the escape route. Change in distance from the forward-looking head marker (nose) to the speaker or the door during signaled active avoidance trials. Negative indicates that the distance toward the noted location was reduced during the orienting response. Shown are all trials, and those trials classified as avoids or escapes. *p < 0.05, significant change.
CS-evoked orienting responses are associated with superior colliculus activation
The superior colliculus has long been considered a central locus for orienting movements (Sprague, 1966; Wurtz and Albano, 1980; Sparks, 1986; Dean et al., 1989). However, its role in CS-evoked orienting responses is not known. In a group of mice (n = 5), we expressed GCaMP6f in the superior colliculus by injecting an AAV with a CaMKII promoter (Chen et al., 2013) known to target output glutamatergic superior colliculus neurons (e.g., Solie et al., 2022). A single optical fiber was implanted in the injection site to monitor neural activity by imaging calcium signals with fiber photometry, as previously described (Hormigo et al., 2021b). Figure 6A shows an optical fiber track in the superior colliculus and GCaMP-expressing neurons measured in the intermediate/deep layers. It is estimated that the optical fiber images a volume (2.5 × 107 mm3) that extends ∼200 mm from its ending (Pisanello et al., 2019), which includes the intermediate/deep layers of the superior colliculus. We measured calcium fluorescence (F/Fo) associated with CS-evoked orienting responses during signaled avoidance procedures (Fig. 6). We did not attempt to compare avoidance procedures or sessions because differences between sessions could be caused by variations in the light path connection.
CS-evoked orienting responses are associated with superior colliculus activation measured with calcium imaging fiber photometry. A, Parasagittal section showing the optical fiber tract reaching superior colliculus and GCaMP6f fluorescence expressed in CaMKII neurons around the fiber ending. The main panel blends a dark-field image of the section with the green channel of the GCaMP6f fluorescent image. Inset, Close-up of the fluorescent image (green and blue DAPI channels) without blending with the dark-field image. B, Cross-correlation between head movement and superior colliculus F/Fo for the rotational and translational components. C, Linear fit (correlation, r) between overall head movement and superior colliculus F/Fo, including the rotational and translational components. D, Traces of F/Fo and head movement rotational and translational components evoked by the CS during AA1/AA2 (left panels) and CS1/CS2 trials during AA3. In these mice, the CS2 tone had a lower SPL (60-65 dB) than the regular CS2 tone, and did not evoke a noticeable inhibitory component. E, Per trial correlations between orienting response overall, rotational, and translational movements with superior colliculus F/Fo. Lines indicate the linear fits between the measured variables before (red) and after scrambling one of them (gray). The correlation of the red line fit is significant (p < 0.0001), but the gray line is not.
To establish the relation between the head movement and F/Fo, we calculated the cross-correlation between these continuous variables per behavioral AA1-AA3 sessions (Fig. 6B), which revealed that both rotational and translational head movements were strongly cross-correlated with superior colliculus F/Fo activation. We then calculated the linear fit between the movement and F/Fo by integrating over a 200 ms window, which revealed a strong linear relationship (correlation; p < 0.05 for all sessions) between superior colliculus activation and overall, rotational, and translational movements (Fig. 6C). This relation was absent when one of the variables was scrambled (Fig. 6C, open symbols). These results indicate that movement during the avoidance procedures is closely related to superior colliculus activation. This is reflected in the overlapping CS-evoked movement and F/Fo traces for the AA1/AA2 and AA3 procedures (Fig. 6D). CS-evoked orienting responses during signaled active avoidance were associated with a sharp F/Fo activation of superior colliculus neurons, which was followed by activation during the ensuing avoidance movement (Fig. 6D). These results are similar to previous unit recordings in the superior colliculus of rats performing signaled active avoidance, which showed sharp firing at CS onset followed by subsequent firing during the avoidance movement (Cohen and Castro-Alamancos, 2010c). Thus, both the orienting and the ensuing avoidance movements were associated with superior colliculus activation. In AA3, the CS2-evoked superior colliculus activation was small in association with the absent orienting response to this less salient stimulus.
We then correlated the overall, rotational, and translational orienting response movements (using the mean of the 0-0.5 s epoch speed per trial) with the superior colliculus activation per each trial in the AA1-AA3 procedures. This revealed a significant correlation (r = ∼0.2; p < 0.05; Fig. 6E, red line and black dots), which was absent when one of the variables was scrambled (gray line). Thus, the orienting response is related to superior colliculus activation. However, this relationship was weaker than the overall movement, which includes the avoidance and escape components (Fig. 6B–D). As per neural recordings (Cohen and Castro-Alamancos, 2010c), there were strong correlations between the avoidance (r = 0.5, p < 0.0001) and escape (r = 0.77, p < 0.0001) movements with the superior colliculus F/Fo activation during the avoidance and escape windows. In conclusion, superior colliculus activation is positively associated with CS-evoked orienting responses, and this relationship is stronger with the larger movements of avoidance and escape responses.
Optogenetic superior colliculus modulation affects CS-evoked orienting responses
Since superior colliculus neurons activate in association with CS-evoked orienting responses, we determined the effects of bilaterally manipulating the activity of superior colliculus neurons on CS-evoked orienting. We used multiple optogenetic approaches to either inhibit (Fig. 7A) or excite the superior colliculus (Fig. 7B), which we previously validated with slice and in vivo recordings (Hormigo et al., 2016, 2019, 2021a,c). We used bilateral manipulations because the auditory CS is also delivered binaurally. The optogenetic groups were compared with themselves (control vs optogenetic trials in the same sessions) and to a No-opsin group (No Opsin, n = 7) that did not express any opsins but was subjected to the same light protocols. As in our previous studies, depending on the opsin expressed, we used blue light as continuous pulses and high frequency (40-66 Hz) trains of 1 ms pulses, or green light as continuous pulses. Light was delivered bilaterally during the avoidance and escape intervals (until the animal avoids or escapes).
Schematic of the different optogenetic groups of mice used in the present study to inhibit (A) or excite (B) the superior colliculus. Blue represents blue light and expression of ChR2. Green represents green light and expression of Arch. As noted, opsin expression occurs in GABAergic (Vgat), glutamatergic (Vglut or CaMKII), or cholinergic (Chat) afferents and/or local neurons in the superior colliculus. The GABAergic afferents originated in SNr. Light was always applied by optical fibers implanted within the superior colliculus. C, PSTHs represent unit firing in superior colliculus (SC) recorded from an implanted optrode in freely moving Vgat-ChR2 mice. Application of blue light (500 ms) into superior colliculus inhibited spontaneous firing. The inhibition was stronger for continuous blue light than for trains at 20 or 40 Hz (10 ms pulses in the trains). D, Population analysis of the data shown in C.
In general, when blue light is delivered to ChR2-expressing cell bodies, firing rate increases linearly with train frequency (1 ms pulses) and is maximal during continuous pulses (i.e., continuous light produces higher firing rates than trains) (e.g., Hormigo et al., 2016). The opposite is the case when light is delivered to ChR2-expressing fibers that cause postsynaptic actions. For example, as we showed previously in Vgat-SNr-ChR2 mice, high-frequency trains of blue light (1 ms pulses) delivered to ChR2-expressing GABAergic SNr afferents inhibit postsynaptic neurons in superior colliculus more efficiently than continuous pulses because the trains causes a more sustained postsynaptic inhibition (Hormigo et al., 2019). In contrast, in Vgat-ChR2 mice, which express ChR2 in local GABAergic superior colliculus neurons, continuous blue light produces stronger inhibition. Indeed, Figure 7C shows the effect of blue light applied using an implanted optrode in the superior colliculus of freely moving Vgat-ChR2 mice (18 sessions in 2 mice) on superior colliculus firing. Continuous blue light (500 ms) strongly inhibited superior colliculus neuronal firing compared with spontaneous firing (Tukey t(51) = 39.8, p < 0.0001, Cont vs Spon; Fig. 7D). While trains of blue light also inhibited neuronal firing (p < 0.0001, 20 Hz or 40 Hz vs Spon), the inhibition caused by continuous blue light was significantly stronger than that caused by either 20 Hz (Tukey t(51) = 16.2, p < 0.0001, Cont vs 20 Hz) or 40 Hz (Tukey t(51) = 26.9, p < 0.0001, Cont vs 20 Hz) trains, although long pulses (10 ms) were used in the trains.
We first determined the effect of inhibiting superior colliculus on CS-evoked orienting by comparing randomly presented AA1 trials (CS) and the same trials including optogenetic light (during the avoidance/escape intervals) to inhibit superior colliculus (CS+Inhibit SC; n = 18 mice, 7 sessions each; Fig. 8A). The data from the different superior colliculus inhibition groups and light stimuli were combined after determining that they produced similar effects on CS-evoked orienting. Inhibition of the superior colliculus had little effect on avoidance performance. It did not affect avoidance latency or the number of ITCs, but it did slightly suppress (90% vs 83%) the percentage of avoidance responses (Tukey t(119) = 7.1, p < 0.0001; Fig. 8A, bottom). At the same time, inhibition of superior colliculus enhanced overall CS-evoked orienting responses during signaled active avoidance (Tukey t(63) = 7.89, p < 0.0001), including both the rotational (Tukey t(63) = 7.46, p < 0.0001) and translational (Tukey t(63) = 7.31, p < 0.0001) movements. Thus, inhibition of superior colliculus has minor effects on the ability of the natural auditory CS to produce active avoidance, and this is associated with enhanced CS-evoked orienting responses. None of these effects on CS-evoked orienting responses occurred in No Opsin mice (n = 9) subjected to the same optogenetics procedures (Tukey t(8) = 0.55, p = 0.7).
Altering CS-evoked superior colliculus activation with optogenetic inhibition or excitation enhances CS-evoked orienting responses during AA1. A, Optogenetic inhibition of the superior colliculus during the CS presentation enhanced overall, rotational, and translational movements of CS-evoked orienting responses compared with CS alone trials. Bottom, Behavioral performance. *p < 0.05, optogenetic versus natural CS trials within the same sessions. B, When optogenetic inhibition of the superior colliculus was used as a CS (without the natural tone CS), the animals sustained high levels of avoidance responses (albeit at lower levels than the natural CS) and the orienting responses evoked by inhibiting the superior colliculus were slightly larger in overall, rotational, and translational movement compared with the orienting responses evoked by the natural CS. C, D, Same as per A, B, but the optogenetic manipulations excited superior colliculus neurons. The orienting responses evoked by the optogenetic stimulation were larger than the orienting responses evoked by the natural CS. However, the enhancement of orienting responses evoked by exciting superior colliculus involved mostly translational, not rotational, components. Although superior colliculus excitation was an effective signal to avoid the US, performance was worse than when the natural CS was used. E, Example orienting response traces showing rotational (red and dashed black) and translational (blue and dashed cyan) movements for the procedure in B. Dashed traces represent optogenetic inhibition of the superior colliculus used as a CS. Continuous traces represent the natural auditory CS. Note the larger dashed traces. F, Example orienting response traces represent rotational (red and dashed black) and translational (blue and dashed cyan) movements for the procedure in D. Dashed traces represent optogenetic excitation of the superior colliculus used as a CS. Continuous traces represent the natural auditory CS. Note the larger translational orienting movement during optogenetic excitation (dashed cyan) compared with the natural CS (blue). Left, The traces on the right expanded in both axes. This reveals a shift of the translational avoidance movement to the left (sooner) when it is evoked by superior colliculus excitation (dashed cyan) compared with the natural CS (blue).
We also determined whether inhibiting the superior colliculus could serve as a CS (without the natural auditory CS) in the AA1 task to evoke orienting responses by comparing randomly presented AA1 trials (CS) and trials that included the optogenetic stimulation alone as the CS (Inhibit SC [as CS]; n = 18 mice, 7 sessions each; Fig. 8B). When superior colliculus inhibition was used as a CS in the AA1 task (without the auditory tone), there was a difference in the percentage of avoids (84% vs 70%) compared with the natural auditory CS (Tukey t(125) = 10.73, p < 0.0001), but not in avoidance latency or ITCs, indicating that superior colliculus inhibition is an effective stimulus to drive avoids, but it is less effective than the natural CS (Fig. 8B, bottom). Moreover, the CS-evoked orienting response increased between the natural auditory CS and the superior colliculus inhibition used as a CS (Tukey t(63) = 4.22, p = 0.004), including both the rotational (Tukey t(63) = 4.6, p = 0.001) and translational (Tukey t(63) = 3.18, p = 0.02) movements. Figure 8E shows example CS-evoked traces from this procedure. The traces show both the translational (blue vs dashed cyan) and rotational components (red vs dashed black) evoked by natural auditory CS or the superior colliculus optogenetic inhibition (dashed traces denote inhibition). Note the larger orienting responses evoked by inhibition than by the natural auditory CS. Thus, superior colliculus inhibition was effective at substituting for the natural auditory CS while evoking a larger orienting response.
We next determined how exciting the superior colliculus affects CS-evoked orienting by comparing randomly presented AA1 trials (CS) and the same trials including optogenetic stimulation (during the avoidance/escape intervals) to excite the superior colliculus (CS+Excite SC; n = 15 mice, 7 sessions each; Fig. 8C). Excitation of the superior colliculus did not affect avoidance performance; however, superior colliculus excitation enhanced CS-evoked orienting responses (Tukey t(98) = 5.06, p < 0.0001), including primarily translational (Tukey t(98) = 5.27, p < 0.0001), not rotational (Tukey t(98) = 3.3, p = 0.02) movements. When superior colliculus excitation was used as a CS in AA1 (without the auditory tone), there was also a difference in the percentage of avoids (95% vs 72%) compared with the natural auditory CS (Tukey t(98) = 12.50, p < 0.0001), but not in avoidance latency or ITCs, indicating that superior colliculus excitation is an effective stimulus to drive avoids (Fig. 8D, bottom). Moreover, the CS-evoked orienting response was slightly larger compared with the natural auditory CS (Tukey t(98) = 3.4, p = 0.01), including primarily translational (Tukey t(104) = 3.74, p = 0.0096), not rotational (Tukey t(98) = 1.1, p = 0.4) movements. Figure 8F shows example traces from this procedure. Note the larger orienting responses evoked during superior colliculus optogenetic excitation (dashed traces denote excitation) but only for the translational component (blue vs dashed cyan), not the rotational component (red vs dashed black). Moreover, the expanded traces (Fig. 8F, right) reveal that the avoidance translational movement shifted leftward (blue vs dashed cyan). This indicates that the avoidance movement started sooner when driven by superior colliculus excitation, although this did not translate into significantly faster avoidance latencies. Thus, superior colliculus excitation was effective at substituting for the natural auditory CS while evoking a larger orienting response involving increased translational movement.
The previous results indicate that both inhibition and excitation of the superior colliculus enhance CS-evoked orienting responses, which suggests a modulatory role of the superior colliculus in regulating these responses at downstream areas. Intriguingly, while inhibition of superior colliculus affected both rotational and translational movements of the CS-evoked orienting response, excitation of superior colliculus affected mostly the translational movement. This suggests that bilateral superior colliculus excitation may directly drive the avoidance response, which involves mostly translational movement, perhaps because there is no need for evaluation of the CS since the colliculus is directly getting the signal to avoid. Alternatively, the bilateral nature of the excitation drives only the translational component of the orienting response, which implies that this component is caused by the coherent activation of both superior colliculi bilaterally, while the rotational component would be driven by unequal activation of both colliculi.
Superior colliculus neuronal firing predicts exploratory orienting direction
The previous results indicate that CS-evoked orienting responses are associated with a robust superior colliculus activation, and that both excitation or inhibition of superior colliculus delivered bilaterally drive the orienting response in the context of signaled avoidance. Next, in a group of mice (n = 5), we recorded calcium imaging fiber photometry from superior colliculus neurons in relation to spontaneous orienting movements during active exploration of an arena (Fig. 9). In another group of mice (n = 5 mice), we implanted sharp tungsten electrodes in the intermediate/deep layers on one side of the superior colliculus to record unit firing during active exploration of an arena (Fig. 10). Using the head tracking, we used an algorithm to identify spontaneous orienting (rotational) movements (contraversive and ipsiversive vs the recording side) and generated peristimulus time histograms (PSTHs) around the movements.
Population superior colliculus neuron activity measured with calcium imaging fiber photometry reveals the direction of spontaneous exploratory orienting movements. A, F/Fo calcium imaging, head turn bias, and head speed traces classified by the turning direction (ipsiversive and contraversive) versus the side of the recording (implanted optical fiber). At time 0, the animals spontaneously turn the head in the indicated direction. Columns represent all head turns (left), those that included a previous turn within 1 s (middle), and those that were devoid of a previous turn within 1 s (right). The speeds of the movements were similar in both directions. B, Population measures (peak amplitude and time to peak) of the F/Fo signal before (pre-) and after (post-) movement onset (zero in A) for >1 s turns (to reduce the impact of previous turns). The negative valley was measured around movement onset. *p < 0.05, between both directions.
Superior colliculus unit activity measured with electrophysiology reveals two groups of neurons that predict spontaneous exploratory orienting movements. A, PSTH of unit firing (Hz), head turn bias, and speed traces classified by the turning direction (ipsiversive and contraversive) versus the side of the recording microelectrodes for >1 s turns. At time 0, the animals spontaneously turned their head in the indicated direction. The speeds of the movements were similar in both directions. B, Units shown in A classified with principal component analysis as Group 1 (75% of units) and Group 2 (25% of units) have two different firing patterns during the movements. C, Population measures of unit firing for Group 1 and Group 2 neurons. Bottom, The firing corrected by the baseline firing. *p < 0.05, significant differences between both directions. #p < 0.05, significant differences between both groups of units. D, A selected group of very well-isolated single units (n = 17) from the population of recorded units all fall within the Group 1 category.
Figure 9A shows superior colliculus F/Fo activation in relation to spontaneous orienting movements (170 sessions in 5 mice with >30 turns in each direction detected per session). Contraversive orienting (black traces) was associated with a large increase in F/Fo activation that tracks the detected movement from onset (time 0), and there was little F/Fo activation before onset. In contrast, ipsiversive orienting (measured within the same sessions) was associated with some F/Fo activation before the onset of the detected movement, but this was sharply curtailed (inhibited) around movement onset as if reflecting a deactivation. Thereafter, ipsiversive orienting was associated with F/Fo activation that was equivalent but slower (delayed) than the F/Fo activation during contraversive orienting. The small ipsiversive F/Fo activation before movement onset appears to be because of the tendency of the animals to be moving in the opposite direction before a sharp orienting turn. Thus, when the detected movement is ipsiversive, the animals may have been moving in the contraversive direction before zero, which causes the small F/Fo activation. Moreover, when the detected movement is contraversive, the animals may have been moving in the contraversive direction before zero, but this does not cause an evident F/Fo activation, which implies that it is the contraversive orienting that is primarily associated with superior colliculus activation.
To reduce the influence of prior turns on the detected movement, we separated the detected orienting movements into those that were free of detected orienting movements in any direction within the preceding second (Fig. 9A >1 s turns) from those that were affected by prior detected movements (Fig. 9A <1 s turns). This revealed that the detected movements that were affected by recent prior detected movements had large F/Fo activations preceding the orienting movement onset, which indicates that the F/Fo activation before detected movement onset reflects the effects of previous turns. Therefore, we used >1 s turns for statistical comparisons, which revealed several differences between the contraversive and ipsiversive orienting activations. First, a sharp F/Fo deactivation around the movement onset only occurs during ipsiversive orienting. This was further estimated by measuring the amplitude of the F/Fo negative valley around the movement onset, which was larger for ipsiversive movements (Fig. 9B, Tukey t(169) = 5.16, p < 0.0001). Second, while orienting movements in both directions produced F/Fo activation starting at movement onset, the contraversive activation rose and peaked faster than the ipsiversive activation (Fig. 9B, Tukey t(169) = 4.7, p < 0.0001) despite the opposite movements being virtually identical in speed and angle. Finally, the peak amplitude of activation was much stronger after movement onset than before movement onset for both ipsiversive (Fig. 9B, Tukey t(169) = 9.8, p < 0.0001) and contraversive (Fig. 9B, Tukey t(169) = 13.2, p < 0.0001) orienting, but there was no difference in amplitude between ipsiversive and contraversive F/Fo activations. Since fiber photometry is a population measure that sums all imaging signals, with limited temporal and spatial resolution, we used unit recordings to decipher superior colliculus neuron activity during spontaneous exploratory orienting movements.
Figure 10 shows PSTHs of superior colliculus unit recordings (n = 89 units from n = 5 mice) during spontaneous exploratory orienting movements (>1 s turns). These units (60.7% of the total) were selected because they showed statistically significant modulation (Wilcoxon Signed Rank test p < 0.05) during detected spontaneous orienting movements in any direction compared with their baseline firing. When these neurons were combined together, contraversive movements were associated with a sharp activation around movement onset (Fig. 10A, black traces). The activity ramped before movement onset and peaked around the movement onset. In contrast, ipsiversive movements (within the same sessions; Fig. 10A, red traces) revealed an inhibitory deactivation around the movement, which begins before movement onset and is interrupted by a sharp activation that was delayed and smaller in amplitude compared with the contraversive activation. We classified the recorded activities by subtracting the paired PSTHs for contraversive and ipsiversive movements and applying principal component analysis to the resulting PSTH trace. Using the four principal components from these traces, we sorted and classified all the recordings into two groups (Fig. 10B). Group 1 included the majority of recorded units (75%), which activated sharply during contraversive orienting and were inhibited during ipsiversive orienting. In contrast, Group 2 included the remaining units (25%), which activated during the movements in both directions, albeit this activation was weaker than for Group 1 cells. Group 2 unit activities also showed a difference in baseline firing before movement onset, which was larger for contraversive movements. Thus, when we measured firing 500 ms around the movement for these groups, without baseline correction (Fig. 10C, top), ipsiversive movement showed less firing in Group 1 but not in Group 2, and contraversive activation was larger in Group 1 than Group 2. However, this difference was absent for ipsiversive movements. In contrast, the same comparisons performed after baseline correction (Fig. 10C, bottom) revealed that Group 1 cells were excited during contraversive movements and inhibited during ipsiversive movements, while Group 2 activities were excited above baseline only during ipsiversive movements. From the population of recordings, we selected the best-isolated single units (n = 17; Fig. 10D). These neurons were all classified into Group 1, displaying pure excitation during contraversive movements and pure inhibition during ipsiversive movements.
Together, the results indicate that most neurons in the intermediate/deep layers of the superior colliculus reliably code the ipsiversive versus contraversive directions of spontaneous exploratory orienting movements. However, a minority of neurons discharge when orienting movements occur in either direction. These neurons may signal the onset, not the direction of an orienting movement. If the different neuronal populations discerned with unit recordings both express CaMKII, their activations will sum in our fiber photometry F/Fo measures, which would explain the occurrence of F/Fo activation for both ipsiversive and contraversive movements in these measures. These results support the well-established role of superior colliculus in coding spontaneous exploratory (directional) orienting movements, which serves as a substrate for the sharp activation of superior colliculus neurons during CS-evoked orienting responses.
Unilateral superior colliculus inhibition produces ipsiversive exploratory orienting
Since superior colliculus neurons activate in association with CS-evoked orienting responses, code the direction of spontaneous exploratory orienting movements, and bilateral excitation and inhibition of superior colliculus both drive orienting responses, we determined the effects of manipulating the activity of superior colliculus neurons unilaterally on spontaneous exploratory orienting movements using the same optogenetic approaches (Fig. 7) used during active avoidance (Fig. 8). The optogenetic groups were compared with a No-opsin group (No Opsin, n = 7) that did not express any opsins but was subjected to the same light protocols. In general, application of continuous (Cont) blue or green light, or trains (1 ms pulses) of blue light at 10-20 Hz or 40-66 Hz in the superior colliculus of No Opsin mice (27 sessions in 7 mice; blue and green light are combined in Cont) while they explored an open field did not affect head direction bias or speed (Fig. 11A,B).
Optogenetic inhibition of the superior colliculus biases ipsiversively spontaneous exploratory orienting movements. A, Each panel overlays the effects of trains (blue only) or continuous (blue or green) light on head bias (angle in degrees), and rotational and translational movement components (speed, cm/s) during exploration of an open field for the different groups used to inhibit the superior colliculus (noted in Fig. 7A), including No Opsin controls. B, Population measures of peak head bias and peak rotational and translational movements for the different groups in A. *p < 0.05 versus No Opsin mice. #p < 0.05 between optogenetic stimuli within a group.
While previous studies have explored the effects of electrical or optogenetic excitation of glutamatergic output cells in superior colliculus (Dean et al., 1986; Masullo et al., 2019; Isa et al., 2020), less is known about the effects of superior colliculus inhibition on orienting. Inhibition of superior colliculus in Vgat-SNr-ChR2 mice by exciting with blue light ChR2-expressing GABAergic superior colliculus afferents originating in the SNr (17 sessions in 4 mice; Fig. 11A,B) with Cont, 10-20 Hz, or 40-66 Hz blue light evoked an ipsiversive orienting head bias (rotational movement) compared with No Opsin controls (Cont vs No Opsin Tukey t(128) = 4.97, p = 0.0078 [bootstrapping p < 0.0001]; 10-20 Hz vs No Opsin Tukey t(128) = 3.98, p = 0.061 [p = 0.003]; 40-66 Hz vs No Opsin Tukey t(128) = 5.02, p = 0.007 [p = 0.003]). The ipsiversive head bias evoked by 40-66 Hz was larger than the bias evoked by 10-20 Hz (Tukey t(52) = 12.89, p < 0.0001 [p < 0.0001]) and Cont (Tukey t(52) = 12.1, p < 0.0001 [p < 0.0001]). Interestingly, the orienting movement (change in direction) occurred at light stimulus onset for trains, but at light offset for Cont (Fig. 11A). However, the movement vigor (overall speed) increased at light onset for both trains and Cont stimuli, and the change in vigor was transient for trains but more sustained for Cont, and these effects involved primarily rotational movement (Fig. 11B). These differences likely reflect the distinct postsynaptic effects of trains and Cont blue light applied to GABAergic synaptic terminals versus cell bodies that express ChR2 (Hormigo et al., 2019); trains produce sustained postsynaptic inhibition, while Cont produces sharp inhibition at stimulus onset that adapts thereafter. In addition, while excitation of GABAergic superior colliculus afferents inhibits superior colliculus neurons, it may also inhibit nuclei targeted by the collaterals of these afferents (through antidromic activation), including the SNr itself, which can cause complex orienting effects (Hormigo et al., 2021a). Thus, we next determined the effects of local superior colliculus inhibition.
Inhibition of superior colliculus neurons by exciting with blue light local ChR2-expressing GABAergic superior colliculus neurons in Vgat-SC-ChR2 mice (72 sessions in 6 mice; Fig. 11A,B) with Cont, 10-20 Hz, and 40-66 Hz blue light evoked an ipsiversive head orienting bias compared with No Opsin controls (Cont vs No Opsin Tukey t(246) = 15.87, p < 0.0001 [p < 0.0001]; 10-20Hz vs No Opsin Tukey t(246) = 5.46, p = 0.0019 [p < 0.0001]; 40-66 Hz vs No Opsin Tukey t(246) = 10.39, p < 0.0001 [p < 0.0001]). Similarly, exciting GABAergic superior colliculus neurons in Vgat-ChR2 mice (26 sessions in 3 mice; Fig. 11A,B) with Cont blue light evoked an ipsiversive head orienting bias compared with No Opsin controls (Cont vs No Opsin Tukey t(110) = 11.55, p < 0.0001 [p < 0.0001]). Since Cont is more effective than 1 ms pulse trains at driving ChR2-expressing cell bodies (Hormigo et al., 2019), the strongest ipsiversive movement was produced by Cont (see also Fig. 7C,D). The change in movement vigor was stronger and persistent during Cont compared with trains, which was weaker and transient, and these effects involved both rotational and translational movement (Fig. 11B).
Inhibition of Arch-expressing glutamatergic superior colliculus neurons in Vglut-Arch mice (14 sessions in 3 mice; Fig. 11A,B) with Cont green light evoked a strong ipsiversive head orienting bias compared with No Opsin controls (Cont vs No Opsin Tukey t(55) = 20.88, p < 0.0001 [p < 0.0001]). There was an increase in movement vigor that was sustained during the light stimulus, and this involved both rotational and translational movement (Fig. 11B). The effects in CaMKII-SC-Arch mice (18 sessions in 3 mice; Fig. 11A,B) were similar to those in Vglut-Arch mice, but the head bias was less prominent, perhaps highlighting a distinction between CaMKII and Vglut neurons in superior colliculus (Fig. 11B).
In conclusion, superior colliculus inhibition with a variety of optogenetic methods produces ipsiversive head orienting, but the dynamics of the ipsiversive orienting movement depend on the characteristics of the optogenetic manipulation.
Unilateral superior colliculus excitation produces contraversive exploratory orienting
Next, we tested the effect of unilaterally exciting the superior colliculus on exploratory orienting bias by using three complementary approaches. Excitation of ChR2-expressing CaMKII superior colliculus neurons in CaMKII-SC-ChR2 mice (46 sessions in 6 mice; Fig. 12) with Cont and 40-66 Hz blue light evoked a contraversive orienting head bias compared with No Opsin controls (Cont vs No Opsin Tukey t(178) = 23.21, p < 0.0001 [p < 0.0001]; 40-66 Hz vs No Opsin Tukey t(178) = 7.19, p < 0.0001 [p < 0.0001]). The strongest movement was caused by Cont blue light, which produced a sharp change in movement vigor at light stimulus onset that was sustained during the stimulus. While the sharp movement at onset was both rotational and translational, the sustained movement was mainly translational and ramping as the animals move forward increasingly faster.
Optogenetic excitation of the superior colliculus biases contraversively spontaneous exploratory orienting movements. A, Each panel overlays the effects of trains (blue only) or continuous (blue or green) light on head bias (angle in degrees), and rotational and translational movement components (speed, cm/s) during exploration of an open field for the different groups used to excite the superior colliculus (noted in Fig. 7B). B, Population measures of peak head bias and peak rotational and translational movements for the different groups in A. *p < 0.05 versus No Opsin mice (shown in Fig. 11). #p < 0.05 between optogenetic stimuli within a group.
Excitation of ChR2-expressing cholinergic neurons and afferents, which activates superior colliculus neurons (Bezdudnaya and Castro-Alamancos, 2014), in Chat-ChR2 mice (18 sessions in 3 mice; Fig. 12) with Cont blue light evoked a contraversive orienting head bias (Cont vs No Opsin Tukey t(84) = 8.51, p < 0.0001 [p < 0.0001]). This effect was modest compared with direct excitation of principal cells and did not cause a change in speed but biased the ongoing movement in the contraversive direction.
Inhibition of Arch-expressing GABAergic superior colliculus neurons, which disinhibits other superior colliculus neurons, in Vgat-Arch mice (25 sessions in 3 mice; Fig. 12) with Cont green light evoked a contraversive orienting head bias (Cont vs No Opsin Tukey t(64) = 10.69, p < 0.0001 [p < 0.0001]). This effect was also accompanied by little change in movement vigor, but it involved mainly rotational movement.
In conclusion, superior colliculus excitation with a variety of optogenetic approaches produces a contraversive head orienting bias.
Discussion
There are several main findings from this study. First, auditory CS-evoked orienting responses consist of a rapid head movement, containing both rotational and translational components, that are sensitive to stimulus intensity. Second, CS-evoked orienting responses varied substantially in amplitude as a function of the behavioral state of the animal during signaled avoidance behaviors; they were largest during uncertain or demanding contexts, such as learning signaled active avoidance or having to discriminate between different CS, respectively. Third, CS-evoked orienting responses did not vary as a direct result of changes in the meaning of the CS. However, a CS that signaled passive avoidance, which signals the need to inhibit/postpone an action, could evoke an inhibitory orienting response by which the animal rapidly suppresses its ongoing head movement, indicating that the orienting response may prepare the animal for the upcoming action. Indeed, during signaled active avoidance trials, the CS-evoked orienting response on average moved the head toward the escape route (door), in an apparent preparation for the upcoming action. Fourth, failed active and passive avoidance responses were associated with larger CS-evoked orienting responses, perhaps reflecting a negative behavioral state usually associated with poor performance. Fifth, while the superior colliculus displayed a well-established role in coding and controlling the direction of spontaneous head orienting during active exploration of the environment, bilateral excitation or inhibition of superior colliculus neural activity both enhanced CS-evoked orienting responses, indicating that these responses are generated downstream but can be modulated by the superior colliculus. Moreover, while bilateral superior colliculus inhibition enhanced both rotational and translational movement components associated with CS-evoked orienting responses, excitation primarily enhanced the translational components, perhaps because of the prominent role of superior colliculus firing in driving signaled avoidance (which is mostly translational, forward movement) via the midbrain tegmentum (Hormigo et al., 2019).
CS-evoked orienting responses
Presentation of a stimulus activates well-known ascending sensory pathways that sequentially engage nuclei in the brainstem, midbrain, and forebrain. In the auditory system, a complex brainstem network feeds auditory inputs to the midbrain, from where activity is relayed to the forebrain. While these pathways have been extensively studied in conditions when animals do not move, much less is known about their role in driving orienting responses. Nevertheless, auditory-evoked startle responses have been extensively studied in behaving rodents, and these extremely fast responses (∼5 ms latency) are usually distinguished from orienting responses because they are only elicited by very high-intensity sounds. Startle responses are typically evoked by loud, white noise pips (>90 dB, <50 ms), compared with the weaker tones used in the present study. Startle stimuli appear to engage a specialized loop involving cochlear root neurons embedded in the auditory nerve, the ventrolateral part of the pontine caudalis nucleus reticularis, and spinal motoneurons (Davis et al., 1982; Lee et al., 1996; Lopez et al., 1999). The extent to which orienting responses engage this startle circuit is unknown, but pontine caudalis nucleus reticularis is targeted by the superior colliculus, which appears to have a major role in mediating both fear-potentiated startle (Zhao and Davis, 2004) and orienting head movements (Sasaki et al., 2004). Future work should improve the differentiation between startle and orienting responses so that the respective brainstem circuit contributions can be deciphered.
Orienting responses reflect behavioral state
Behavioral states are a complex construct that may be defined according to many overlapping variable dimensions, including vigilance (e.g., sustained attention, alert, quiescent, slow wave and REM sleep), motivation (e.g., hungry, thirsty), and emotion (e.g., calm, stressed, anxious, fearful, depressed). Some of these variables are easy to evaluate using external and neural circuit measures, and they can be reliably adjusted by manipulating the animal's internal or external environment. Other variables, particularly emotional variables, are more difficult to evaluate because they depend on complex interactions between the situation, the animal's assessment of the situation, and its ability to cope with it. During active avoidance training, animals adjust their behavioral state from an initial high alert level to a quiescent state as the training progresses, and these changes are associated with shifts in the operating mode of neural systems, such as the level of activation and dynamics of the sensory thalamocortical pathways processing the CS (Castro-Alamancos, 2004a,b, 2010). These changes are certainly also associated with adjustments in emotional variables that are generally harder to evaluate in animals. For instance, the degree of fear in signaled active avoidance has been discussed extensively (e.g., Mowrer, 1960; Kamin et al., 1963; Mineka, 1979). It would be useful to have additional behavioral (external) measures that vary with these hard to evaluate, complex emotional variables.
We found that CS-evoked orienting responses varied substantially in amplitude as a function of the behavioral state of the animal imposed by different signaled avoidance contingencies. CS-evoked orienting responses were largest during the initial AA1 learning session, when mice are highly alert because of the uncertain situation they are being presented. This enhancement habituated rapidly, as the animals developed the skill to cope with the new situation (by avoiding the US). Habituation of the CS-evoked orienting response was accompanied by behavioral signs, such as increased active exploration, that indicate a reduction in the levels of negative emotional variables (e.g., fear, anxiety, or stress). Changes in task contingencies that did not alter the meaning of the CS but increased behavioral signs of these negative emotional variables (decreased active exploration and signaled action latency adjustments) enhanced CS-evoked orienting. Moreover, the enhancement of CS-evoked orienting became persistent during demanding contexts, such as when a discrimination between different CS is required to select the correct action to cope with the situation.
In addition, CS-evoked orienting responses appear to prepare the animal for the appropriate upcoming action signaled by the CS. Thus, a CS that signaled the requirement for an inhibitory action (signaled passive avoidance) could produce an inhibitory orienting response by which the animal rapidly suppresses its ongoing head movement. However, this does not imply a prototypical freezing behavior (i.e., the absence of any movement, except for respiratory-related movements) because mice continue to move during these CS. Moreover, failed active and passive avoidance responses were associated with larger orienting responses, perhaps reflecting a negative state associated with poor performance. Together, the results indicate that CS-evoked orienting responses reflect the behavioral state of the animal set by the demands of the current situation and may prepare the animal for the action signaled by the CS. Stimulus-evoked orienting responses could be a useful probe to assess complex behavioral states resulting from exposure to harmful, threatening, uncertain, or demanding situations (i.e., negative emotional states), which are otherwise difficult to evaluate in rodents.
Our findings are in line with an influential theory on habituation derived from spinal reflexes (Groves and Thompson, 1970; Thompson, 2009), by which a stimulus evokes two independent processes in the CNS: one decremental (stimulus selective and affecting mainly the pathway that mediates the reflex) that causes habituation, and one incremental or amplifying (stimulus-independent and affecting neural systems broadly) that causes sensitization. Related models emphasize the development of a memory trace of the stimulus, which is compared with the stimulus presentation on every trial. If the comparison is equal, it leads to habituation by inhibiting the amplifying, state system; otherwise it activates the state system (Sokolov, 1963; Konorski, 1967; Wagner, 1979). Our results provide a new experimental framework to address the neural substrates underlying these systems in future work.
Role of superior colliculus
In our experiments, the superior colliculus displayed a well-established role in coding and controlling the direction of spontaneous orienting head movements during active exploration of the environment (Dean et al., 1986; Wilson et al., 2018; Masullo et al., 2019; Isa et al., 2020). For instance, unilateral optogenetic excitation generated contraversive orienting, whereas optogenetic inhibition generated ipsiversive orienting. Recordings from superior colliculus neurons showed that firing was associated with orienting direction. Most neurons in the intermediate/deep layers discharged briskly during contraversive orienting, and the same neurons were inhibited during ipsiversive orienting. A minority of neurons discharged in association with head movement regardless of the direction. These neuronal dynamics were better deciphered using unit recordings than fiber photometry, likely because the imaging integrates the signals from many neurons combining the distinct dynamics of different groups of neurons.
Regarding CS-evoked orienting, bilateral optogenetic excitation or inhibition of superior colliculus enhanced orienting responses, indicating that these responses are generated downstream in motor circuits of the medulla (currently being defined in mice) (Arber and Costa, 2022), perhaps driven by direct sensory afferents (including auditory) (Lopez et al., 1999), where they can be modulated by superior colliculus activity. Indeed, our fiber photometry calcium imaging recordings of superior colliculus neurons during signaled avoidance showed activation in association with the CS-evoked orienting responses, which can be the basis for this modulation. Interestingly, this activation correlated positively with CS-evoked orienting responses. While we did not perform unit recordings during signaled avoidance in the present study, our previous unit recordings showed that escapes (avoidance failures) are associated with greater firing during the orienting response period than avoids (Cohen and Castro-Alamancos, 2010c). Superior colliculus is critically involved in mediating signaled active avoidance when transmission of the CS to the forebrain through the thalamus is impaired (Cohen and Castro-Alamancos, 2007, 2010a). Under normal circumstances, superior colliculus seems to have a synergistic (not unique) role in transmitting the CS to the underlying pedunculopontine tegmentum, which is an essential area for signaled avoidance (Hormigo et al., 2019, 2021b). Although the superior colliculus activates during CS-evoked orienting responses, it appears to have a complex modulatory role in adjusting these orienting responses, which are likely generated downstream, where many other brain areas can exert an influence as a function of behavioral states. Certainly, the superior colliculus can be part of a purported sensitizing state system since its modulation potentiates CS-evoked orienting responses.
Footnotes
This work was supported by National Institutes of Health grants R35 NS097272 and R01 NS104810 to M.A.C.-A.
The authors declare no competing financial interests.
- Correspondence should be addressed to Manuel A. Castro-Alamancos at mcastro{at}uchc.edu
