Introduction

Attention deficit hyperactivity disorder (ADHD) affects approximately 5% of children and adolescents globally and often persists into adulthood [1,2,3,4]. Characterized by persistent inattention, hyperactivity, and impulsivity, ADHD significantly impairs cognitive function, educational attainment, and professional success [3, 5,6,7]. Despite extensive research, the neurobiological mechanisms underlying ADHD remain elusive, necessitating further investigation to inform more effective treatments. The optimal stimulation theory posits that cognitive deficits in ADHD arise from chronic hypoarousal, leading individuals to seek external stimulation to achieve optimal cognitive functioning [8, 9]. This hypoarousal manifests as reduced alertness, impairing sustained attention and memory processes [10]. Behavioral interventions utilizing external stimuli, such as auditory and visual cues, have temporarily enhanced arousal levels and cognitive performance, effectively modulating alertness critical for information encoding and learning [11,12,13,14].

Despite the fundamental role of alertness in cognitive functions, the cellular and circuit mechanisms regulating alertness and its impact on cognitive function in ADHD are poorly understood. The hypothalamus, particularly the supramammillary nucleus (SuM), is implicated in arousal regulation [15, 16]. The SuM promotes wakefulness and modulates hippocampal activity essential for learning and memory [17,18,19,20]. Recent studies indicate the SuM’s involvement in contextual encoding and activation in response to novelty [19, 20], suggesting it as an integral node in arousal and cognitive processes relevant to ADHD.

Abnormal hippocampal synaptic plasticity has been implicated in ADHD-related cognitive deficits [20, 21]. The dentate gyrus (DG) plays a crucial role in encoding new information and differentiating similar experiences, essential for effective learning [22]. The SuM–DG circuit may facilitate these functions by influencing synaptic plasticity mechanisms [20, 23]. Optogenetic modulation of SuM neurons enhances hippocampal activity and improves cognitive performance, indicating therapeutic potential in targeting this circuit for ADHD [24, 25].

Here, we hypothesized that SuM-DG circuit activation would enhance alertness and improve cognitive performance. We employed behavior tests including looming, new environment exposure and novel object recognition (NOR) to assess the level of alertness and cognition. We utilized electrophysiological recording, circuit tracing, optogenetic and chemogenetic manipulation to interrogate the anatomy basis and functional roles of SuM-DG circuit in deficits in alertness and recognition in ADHD. Finally, we assessed the roles of synaptic plasticity, specifically LTD, played in SuM-DG functions.

Materials and methods

Animals

Male C57BL/6 mice (10–12 weeks old; Jackson Laboratory) and male Sprague-Dawley and spontaneously hypertensive rats (SHRs) (4 weeks old; Charles River Laboratories) were used in this study. Animals were maintained on a 12-h light/dark cycle with controlled temperature (20–22 °C) and humidity (40–60%). Food and water were provided ad libitum.

In this study, animals were randomly assigned to experimental and control groups to minimize potential biases and ensure the reliability of the results. Randomization was applied during animal selection, behavioral testing, and intervention procedures, including optogenetic and chemogenetic manipulations, to ensure comparable baseline characteristics across groups. The sample size estimate was determined through a priori power analysis based on expected effect sizes from previous studies utilizing similar behavioral and electrophysiological assessments. Using GraphPad Prism, we calculated the minimum number of animals required per group to achieve a statistical power of at least 80% with an alpha level of 0.05. This ensured that the study had sufficient power to detect significant differences while adhering to ethical guidelines for animal research by minimizing unnecessary use of subjects.

Blinding was implemented during behavioral testing, electrophysiological recordings, and histological analyses to minimize bias. Experimenters conducting these assessments were unaware of the group allocations. However, due to the nature of surgical procedures and optogenetic/ chemogenetic manipulations, full blinding during these interventions was not feasible. To mitigate potential bias, data analysis was performed in a blinded manner, with group identities concealed until statistical analyses were completed.

Inclusion/exclusion criteria

This study required that all animals be within the specified age range, free of any observable health issues, and exhibit normal baseline behavioral activity prior to experimental procedures. Animals were excluded from the analysis if they exhibited signs of illness, injury, or abnormal behavior that could confound the results. Additionally, animals were excluded if histological verification revealed inaccurate viral injection sites or fiber optic placements, or if electrophysiological recordings failed to meet quality control criteria. These exclusion criteria were pre-established to ensure the integrity and reliability of the data.

Adeno-associated virus (AAV) injections

Animals were anesthetized with 2.0% isoflurane delivered at a flow rate of 0.6–0.8 L/min and received analgesia via subcutaneous injection of meloxicam (0.5 mg/kg). Animals were positioned in a stereotaxic frame (Kopf Instruments) for precise viral injections and optical fiber implantation. A small craniotomy ( ~ 0.6 mm diameter) was performed above the dentate gyrus (DG) or supramammillary nucleus (SuM). Using a 10 μL Hamilton syringe connected to a microinjection pump (Stoelting), 0.5 μL of AAV was stereotaxically injected at a rate of 0.05 μL/min. The syringe remained in place for 5 min post-injection to prevent backflow before being slowly withdrawn. The incision was sutured, and mice were allowed to recover in their home cages for 3 weeks prior to further experimentation. All AAV injections were bilateral unless specified otherwise.

For retrograde tracing, mice received 0.5 μL of AAV2/Retro-CBA-Cre virus (1 × 1012 genome copies [GC]/mL) into the DG (coordinates: anterior-posterior [AP] –1.8 mm, mediolateral [ML] ±1.1 mm, dorsoventral [DV] 2.0 mm; AP -3.0 mm, ML ±1.3 mm, DV 4.1 mm for rats) and 0.5 μL of AAV5/8-EF1α-DIO-eGFP into the SuM (AP –2.80 mm, ML ± 0.5 mm, DV 4.8 mm; AP −4.4 mm, ML ±0.7 mm, DV 8.8 mm for rats). To specifically express channelrhodopsin-2 (ChR2 [E123A]) or halorhodopsin (NpHR3.0) in SuM neurons projecting to the DG, AAV2/Retro-CBA-Cre injected into the DG and AAV5/8-EF1α-DIO-ChR2 (E123A)-eGFP or AAV5/8-EF1α-DIO-NpHR3.0-eGFP (1 × 1012 GC/mL) injected into the SuM. For chemogenetic modulation, AAV5/8-EF1α-DIO-hM4Di-eGFP or AAV5/8-EF1α-DIO-hM3Dq-eGFP (1 × 1012 GC/mL) was injected into the SuM to express inhibitory or excitatory designer receptors exclusively activated by designer drugs (DREADDs), respectively. Unilateral activation of SuM fibers was performed in the right hemisphere, while bilateral inhibition involved injections into both hemispheres.

Behavior tests

The looming test was performed in the open box (40×60×40 cm for mouse and 60 × 120 × 40 cm for rat) with a shelter nest placed in the corner. A computer monitor was placed on the ceiling of the open field box to present looming stimuli, which was a black disc expanding from a visual angle of 2° to 50° in 0.3 sec in a white background. The day before the formal looming test, the mice or rats were allowed to explore the box for 5 min and make sure the mouse entered the shelter nest and remembered the location. During the adaption, the screen was kept displaying the white background, but no looming stimuli was given. For the formal looming test, when the mouse came to the trigger zone and stood up on its hind limbs or raised its head to look up, a looming stimulus was given. For novel environment exposure, animals were introduced to a novel behavioral box with a novel object from their own home cage.

For NOR, during the training, animals were exposed to an open field box (40×40×40 cm for mouse and 80×80×40 cm for rat) containing two similar objects, each 5 cm/8 cm from the edge of the box. Animals were allowed to freely access and familiarize with the two objects in the box for 5 min. Animals were returned to their home cages after the training phase. The testing session was conducted 24 h after training phase. Animals were introduced to the open field box after replacing one of these two objects into a novel object. Object exploration was operationally defined as the animal directing its nose to within ≤2 cm of the object, with the criterion that the frequency of object investigation must exceed 5 approaches. Object positions were counterbalanced across animals.

Immunofluorescence

The animals were deeply anesthetized with 0.4 ml 1.25% 2, 2, 2-tribromothanol (TG-Avertin-R, 2402 A, Tigergene). Brain tissue was perfused with normal saline followed by 4% paraformaldehyde (PFA) solution, dehydrated with 30% sucrose and frozen sections to obtain 30 μm sections with a cryostat microtome (Leica). After incubation with antibodies against c-Fos (rabbit monoclonal, 1:1,000 dilution, cat. no. 2250S, Cell Signaling Technology) overnight at 4 oC, the brain slices were labelled with secondary antibodies (Alexa Fluor 488, rabbit monoclonal, 1:1000 dilution, cat. no. 4412S, Cell Signaling Technology) to exhibit the expression of c-Fos in the brain. c-Fos positive cells were imaged under Zeiss Axio Zoom Macroscope. For the following c-Fos immunostaining, the animals were sacrificed 1 h after the behavior tests.

Optogenetic and chemogenetic manipulations

Three weeks after AAV injections, Animals underwent optical fiber implantation. Optical fibers (200 μm core diameter; CF230-10, numerical aperture 0.22; Thorlabs) were tested for light transmission efficiency ( > 90% at 473 nm) using an optical power meter (PM20, Thorlabs). Fibers were implanted above the SuM (AP –2.8 mm, ML ± 0.5 mm, DV 4.8 mm for mice; AP −4.4 mm, ML ±0.7 mm, DV 8.8 mm for rats) for behavioral experiments assessing alertness and learning and memory, or above the DG (AP –1.8 mm, ML ± 1.1 mm, DV 2.0 mm) for in vivo electrophysiological recordings and activation of SuM fibers during cognitive testing. The ferrules were secured to the skull using dental cement, and mice were allowed to recover for 5–7 days before optogenetic experiments.

For optogenetic activation of SuM fibers in the DG, 1.8 mW of 473 nm blue light was delivered using a theta-burst stimulation (TBS) protocol: bursts of 4 pulses at 100 Hz, with 200 ms intervals between bursts, repeated 10 times. This stimulation pattern mimics the firing patterns observed in SuM neurons [26, 27]. For optogenetic inhibition, 1.8 mW of 589 nm yellow light was used. Sham controls received the same light stimulation protocol with the light blocked above the skull.

Chemogenetic modulation was achieved by intraperitoneal injection of clozapine N-oxide (CNO; 2 mg/kg; Tocris Bioscience) into mice expressing hM4Di or hM3Dq, administered 30 min prior to behavioral testing.

Peptide treatment

To investigate the role of long-term depression (LTD) in recognition, mice received intraperitoneal injections of the GluA2A-3Y peptide (3 mg/kg, HY-P2259), an inhibitor of AMPA receptor GluA2 subunit endocytosis known to block LTD, or a scrambled control peptide. Peptides were administered 30 min before testing.

Electrophysiological recordings and spike sorting

For patch-clamp recordings in vitro, hippocampal slices (300 μm) were prepared and transferred to artificial cerebrospinal fluid (in mM: 124 NaCl, 3 KCl, 26 NaHCO3, 1.2 MgCl2•6H2O, 1.25 NaH2PO4•2H2O, 10 C6H12O6, and 2 CaCl2, at pH 7.4, 305 mOsm). The slices were allowed to recover at 31.5 degree for 30 min and then at room temperature for 1 h. Acute brain slices were then transferred into recording chamber superfused with oxygenated artificial cerebrospinal fluid continuously (2 ml/min) at room temperature and were visualized via IR-DIC using an Axioskop 2FS equipped with Hamamatsu C2400-07E optics (Hamamatsu City, Japan). Baseline electrophysiological properties were recorded when stable whole-cell recordings were achieved with good access resistance (8 ~ 20 MΩ). To show the direct connections of SuM-DG, the presynaptic neurotransmitter release of patched cells under TTX application, was triggered by 473 nm laser pulses (20 ms duration, 2 mW/mm²) through a 200 µm optical fiber positioned approximately 200-300 µm from the recorded DG neurons. The light evoked EPSCs and IPSCs were recorded using an internal solution (in mM, 140 potassium gluconate, 10 HEPES, 0.2 EGTA, 2 NaCl, 2 MgATP, and 0.3 NaGTP), and an external solution containing 1 μM TTX (Tocris Bioscience) and 5 μM DNQX (MedChemExpress), or 1 μM TTX and 10 μM bicuculline (Tocris Bioscience), respectively.

Electrophysiological recordings in vivo as described before with modifications [27]. An array of eight stereotrodes constructed from twisted 17 μm HM-L-coated platinum–iridium wire (20% platinum; #100-167, California Fine Wire) was connected to a custom microdrive for vertical adjustments. During surgery, the stereotrodes were positioned above the recording site and secured with dental cement, and the depth was documented for subsequent estimation of recording sites in the target brain region. Signals were filtered at 250 Hz high-pass for spikes. Data acquisition and initial processing were conducted using OmniPlex software (Plexon). Spike sorting was performed with Offline Sorter software (Offline Sorter, Plexon) and cluster quality was assessed using isolation distance and L-ratio metrics and Plexon SDK (www.plexon.com/software-downloads/SDK). Only the units with an L-ratio less than 0.1 and a distance more than 15 were counted. We isolated and analyzed spike units from individual neuronal types through calculating the valley-to-peak time and the half-width of the spikes. Recording sites in the DG or SuM were estimated by the depth of electrode advancement through the brain using a custom microdrive with 30-µm incremental steps. Spikes in the DG were validated via optogenetic activation. For behavior-evoked SuM responses, behavioral labels were stored in pl2 files as events for subsequent analysis of event-related firing of SuM neurons.

For in vivo plasticity studies examining the effects of SuM optogenetic stimulation on DG plasticity in free moving mice, two of the previous eight stereotrodes were selected as stimulating electrodes. From the eight circularly arranged electrodes, 2-4 contralateral electrodes (approximately 300 μm away from two stimulating electrodes) were used as recording electrodes. Optogenetic stimulation consisted of one train of the aforementioned TBS pattern at 473 nm with 1.8 mW power. Baseline electrical stimulation intensity was determined as the current that evoked field potentials of approximately 500 μV at the recording electrodes, with a maximum current limit of 100 μA. Plasticity was evaluated by comparing the amplitudes of electrically-evoked field potentials before and after optical stimulation.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 8 software. Data are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism 8 software. Non-behavioral data were analyzed using two-tailed unpaired t-tests. Behavioral data were evaluated using one-way or two-way analysis of variance (ANOVA), followed by appropriate post hoc tests (e.g., Tukey’s multiple comparisons test). A p-value less than 0.05 was considered statistically significant. Statistical comparisons were conducted under the assumption of similar variance across conditions. The homogeneity of variance was checked before performing ANOVA or t-tests, as stated in the Results section. No violations of variance assumptions were reported, ensuring the robustness of statistical comparisons.

Results

ADHD rats exhibit deficits in alertness and recognition

To elucidate the relationship between alertness deficits and cognitive impairments characteristic of ADHD, we employed spontaneously hypertensive rats (SHRs) an established animal model that recapitulates the core behavioral phenotypes of ADHD, including hyperactivity, impulsivity, and attentional deficits. SHRs exhibit neurobiological and neurochemical alterations analogous to those observed in individuals with ADHD, making them a valid model for investigating underlying neural mechanisms [28,29,30]. A key methodological consideration in ADHD animal models is the selection of appropriate control groups. While the spontaneously hypertensive rat (SHR) is widely used as an ADHD model, the Wistar-Kyoto (WKY) rat has historically been used as its control due to shared genetic lineage. However, growing evidence suggests that WKY rats are not ideal controls due to substantial differences in locomotion, exploratory behavior, and response to environmental stimuli. In our systematic comparison of WKY and Sprague-Dawley (SD) rats, we found that in the visual looming test, WKY rats exhibited comparable alertness levels to SD rats (Fig. S1B) but showed significantly reduced response frequency (p < 0.001, Fig. S1A), indicating a profoundly diminished exploratory drive. Additionally, during object vigilance testing, WKY rats matched SD controls in threat detection latency (Fig. S1C) but exhibited virtually absent object exploration (zero-crossing events in 87% of trials, Fig. S1D) and markedly reduced locomotion (42% decrease vs. SD, p < 0.01, Fig. S1E).

These behavioral characteristics align with previous findings highlighting WKY rats’ reduced motivation, hypoactivity, and atypical stress responses, which make them an inappropriate control for SHR behavioral studies [31, 32]. Additionally, SD rats exhibit more stable behavioral profiles than WKY rats and serve as a more appropriate control for SHRs [30]. Given these findings, we selected SD rats as controls for our ADHD model to ensure accurate behavioral comparisons, avoiding confounds introduced by WKY-specific traits unrelated to ADHD phenotypes.

We quantified alertness levels using a visual looming stimulus paradigm designed to evoke innate escape responses by simulating an approaching predator. In this setup, a looming stimulus—an expanding dark circle—was projected onto a screen positioned above the open-field arena when the rat’s head entered a predefined trigger zone (Fig. 1A). The latency to initiate an escape response, measured as the time taken for the rat to retreat to a designated safe area upon stimulus presentation, served as an index of alertness and rapid attentional processing. The ADHD rats exhibited significantly prolonged reaction times during the initial three trials of looming stimulation compared to control Sprague-Dawley (SD) rats (Fig. 1B), suggesting impaired alertness mechanisms when confronted with salient environmental threats.

Fig. 1: Alertness-related pretreatment promotes recognition in ADHD rats.
figure 1

A Schematic timeline of the experimental procedures for looming test and NOR test. B Line graph depicting the average retreat latency in response to looming stimuli across trials for control rats and ADHD rats. ADHD rats exhibit significantly longer retreat latencies compared to control rats. C Above, representative movement trajectories during the NOR training session of ADHD and control rats with or without looming pretreatment. Below, bar graph depicting the average latency to initiate object exploration (sniffing) during the training session. D Above, schematic diagram of the NOR test setup during the test phase, indicating the placement of the familiar and novel objects. Below, bar graph showing the average percentage of time spent exploring the novel object during the test phase for control and ADHD rats with or without looming pretreatment.

Full size image

To assess cognitive function, we employed the novel object recognition (NOR) task, which leverages rodents’ innate preference for exploring novel objects over familiar ones. During the training phase, rats were exposed to two identical objects placed in the testing arena, and the latency to initiate exploration—defined as the time from placement in the arena to the first sniffing of an object—was recorded. Consistent with the looming test, ADHD rats displayed significantly shorter latencies compared to control rats (Fig. 1C), and this baseline alertness in ADHD could be increased by the pretreated looming stimuli (Fig. 1C), indicating alertness impaired in ADHD rats when confronted with potential and possible environmental threats, and the potential for behavioral interventions to improve alertness.

In the testing phase, control rats exhibited a significant preference for the novel object, consistent with intact recognition (Fig. 1D). In contrast, ADHD rats did not display a significant preference, allocating comparable amounts of time to both objects, thereby revealing a pronounced deficit in recognition (Fig. 1D). The looming stimulation enhanced recognition level of ADHD rats in the following NOR testing phase as increasing the preference for the novel object (Fig. 1D), indicating a critical connection between alertness and cognition in ADHD, supporting the validity of behavioral interventions aimed at enhancing alertness and cognitive function.

The SuM mediates alertness

To elucidate the cellular mechanisms underlying the alertness deficits observed in ADHD, we focused on the hypothalamus—a critical brain region integral to the regulation of wakefulness and arousal states. The hypothalamus contains several nuclei involved in sleep-wake cycles, stress responses, and alertness modulation, making it a pivotal area for exploring neural correlates of alertness impairments.

We analyzed c-Fos expression, a well-established marker of neuronal activation, in the hypothalamus of wild-type mice subjected to a visual looming test. Mice were habituated in an open-field arena for three hours before receiving five trials of visual looming stimuli (Fig. 2A). One hour after the final looming trial, mice were sacrificed and the brains were processed for c-Fos immunohistochemistry. Compared to non-stimulated mice, the looming-stimulated group exhibited significantly elevated levels of c-Fos expression in several hypothalamic nuclei, including the anterior hypothalamus (AH), posterior hypothalamus (PH), and the SuM (Fig. 2B and C). Among these regions, the SuM showed the most substantial increase in c-Fos-positive cells following visual looming, suggesting a prominent role in mediating alertness and defensive behaviors in response to salient stimuli. To figure out whether looming stimulations activate SuM neurons, we employed in vivo electrophysical recording to detect neuronal spiking in SuM with visual looming stimuli simultaneously in wild-type mice. We found that visual looming stimulations evoked spiking in SuM significantly with a latency of approximately 0.1 s (Fig. 2D). Together, these results indicate that visual looming stimulation, as a type of risk information inputs, is able to activated SuM neurons.

Fig. 2: Alertness-related tasks elevate neuronal activity in SuM.
figure 2

A Above, timeline of the experimental protocol illustrating the application of looming stimuli. Middle, schematic graph of looming test. Below, average reaction time of WT mouse of 1st to 5th trials of looming test. B Diagrams and corresponding micrographs showing the distribution and expression of c-Fos in SuM with and without exposure to looming stimuli. C Bar graphs quantifying the number of c-Fos-positive cells in various hypothalamic regions including the anterior hypothalamus (AH), lateral hypothalamus (LH), posterior hypothalamus/dorsomedial hypothalamus (PH/DM), and SuM, with or without looming stimulation. D Looming-evoked neuronal spiking in SuM. Above, spiking counts per 10 ms before and after looming stimulation. Below, average spiking counts after looming stimulation. Bar, 2 counts/bin (10 ms). E Above, schematic representation of the experimental paradigm for the novel environment exposure test. Below, time withdraw from sniffing the cue. F Representative micrographs showing increased c-Fos expression in the SuM after exposure to a novel environment compared with control in (B). G Representative micrographs showing c-Fos expression in the SuM of control rats and ADHD rats, both with and without looming stimuli. H Statistical analysis presented as bar graphs, illustrating c-Fos expression across control and ADHD groups with and without looming stimuli.

Full size image

Recognizing that potent stimuli like looming can activate neural circuits associated with fear and defense, we investigated whether the SuM is also responsive to milder, non-threatening stimuli that induce alertness. Mice were exposed to a novel environment known to elicit exploratory behavior and heightened alertness due to unfamiliarity for ~2 min and the time withdraw the cues was increased gradually with sniff times, indicating a process to explore the strange environment (Fig. 2E). c-Fos analysis revealed that exposure to the novel environment significantly increased neuronal activation in the SuM compared to control mice maintained in their home cages (Fig. 2F and Fig. S2). These findings indicate that the SuM is responsive to various forms of alertness-inducing stimuli, extending beyond high-threat conditions.

To further examine the relationship between SuM activity and alertness deficits characteristic of ADHD, we conducted c-Fos analysis in control and ADHD rats with or without looming tests. The ADHD rats exhibited reduced neural activity in the SuM compared to control rats (Fig. 2G and H), which consistent with our previous observations of diminished alertness in ADHD rats. Specifically, baseline c-Fos expression in the SuM was significantly lower in ADHD rats, correlating with their hypoalertness. Notably, when SHRs were subjected to looming stimulation, c-Fos expression in the SuM increased markedly, approaching levels observed in stimulated control rats (Fig. 2G and H). This suggests that external stimuli can effectively activate SuM even in ADHD models, potentially compensating for baseline hypoactivity.

Together, these results underscore the critical role of the SuM as a neural hub integrating various alertness-inducing signals from both high-threat and mild stimuli, thereby influencing arousal states and attentional processes essential for optimal cognitive functioning. The reduced activity of the SuM in SHRs correlates with their observed alertness deficits, implying that hypoactivity within this nucleus may contribute to the pathophysiology of ADHD. Notably, the capacity of looming stimulation to augment SuM activity and restore alertness levels in SHRs highlights the potential of targeting the SuM for therapeutic interventions. Enhancing SuM activity could thus represent a promising avenue for addressing the alertness and attentional deficits prevalent in ADHD, ultimately ameliorating associated cognitive impairments.

The SuM-DG circuit mediates alertness and cognition enhancements in ADHD

Building upon our findings that the supramammillary nucleus (SuM) plays a pivotal role in modulating alertness, we investigated the specific neural circuitry through which the SuM influences cognitive functions. Given that SuM glutamatergic neurons densely innervate the hippocampal dentate gyrus (DG), and c-Fos positive cell number in DG is observed increased after looming stimulations (Fig. S2), we hypothesized that the SuM–DG circuit serves as a critical link mediating alertness-induced enhancement in learning and memory that are notably impaired in ADHD.

To map the connections between SuM neurons and the DG, we employed a Cre-loxP system to selectively express green fluorescent protein (GFP) in SuM neurons that project towards DG. Retrograde and anterograde tracing experiments revealed robust projections from the SuM to the DG, confirming the anatomical basis for direct communication between these regions (Fig. 3A; Fig. S3). During the visual looming test, about 70% c-Fos immunoreactivity was colocalized with GFP in SuM (Fig. 3B and C). Similarly, exposure to a novel environment also resulted in co-labeling of GFP and c-Fos in DG-projecting SuM neurons (Fig. S2B). This co-localization underscores the involvement of the SuM-DG circuit in mediating responses to various forms of alertness-inducing stimuli.

Fig. 3: The SuM-DG circuit mediates response to alertness-related inputs.
figure 3

A Schematic and fluorescence images showing the viral tracing strategy used to label projections from SuM to DG. B Timeline of the experimental procedure and bar graph depicting the percentage of SuM neurons co-expressing c-Fos and GFP. C Representative fluorescence images showing co-localization of GFP (green) and c-Fos (red) in SuM neurons. D Schematic showing in vivo electrophysiological recording and optogenetic activation used to depict functional connections of SuM-DG. E DG-projected SuM Neuronal spiking counts evoked by a transient optogenetic activation at SuM-DG neuronal fibers. F Above, schematic of the viral strategy employed for optogenetic manipulation of SuM neurons. Below, representative fluorescent images showing the AAV injection and expression at bilateral SuM. G Above, timeline of looming tests with optogenetic manipulation. Below, reaction times of animals in response to looming stimuli across 30 trials and under sham light stimulation in 11th to 18th trials. H Reaction times of animals in response to looming stimuli across 30 trials and under blue (activation) or yellow (inhibition) optogenetic manipulations. I Schematic of novel environment exposure experiment with sham, blue, and yellow optogenetic manipulations in the second phase. J Bar graph showing latency to sniff novel objects under sham, blue, and yellow light conditions. K Bar graphs illustrating time withdraw from object before, during, and after the optogenetic activation and inhibition.

Full size image

Considering fast responses could be involved to respond the alertness-inducing information inputs like looming stimulation, we then employed optogenetic manipulation to activate DG neuronal fibers that are projected from SuM via expressing channelrhodopsin-2 (ChR2) in SuM neurons and delivering blue light stimulation into the DG, and using in vivo electrophysiological record at DG neurons to figure out the functional patterns of SuM-DG circuit (Fig. 3D). We found that the spiking of DG generated immediately after the event without obvious latency (Fig. 3E), indicating DG neurons display fast responses to the SuM inputs. Besides, in vitro electrophysiological recording also suggested a robust and fast-respond SuM-DG circuit exists (Fig. S3).

To directly interrogate the functional role of the SuM-DG circuit, we utilized optogenetic techniques to activate or inhibit bilateral SuM projections to the DG selectively to assess the alterations in alertness-associated behavior tests (Fig. 3F). Optogenetic light were given at 11–18 trials of looming tests, and the behavioral results revealed that the activation of SuM-DG circuit significantly enhanced alertness levels given the reaction time to the looming stimulation was decreased when blue light was given, while inhibiting the SuM-DG circuit after the alertness stimulation response reached a plateau demonstrated a small trend toward increased reaction time in looming tests (Fig. 3G and H), suggesting that activating the SuM-DG pathway before reaching the alertness plateau plays a critical role in the alertness phenotype.

We also tested whether optogenetic activation or inhibition alter responsive to alertness-induced non-threatening new environment (Fig. 3I). Activation of SuM-DG circuit significantly increased the latency to sniff the object and decreased the time spending in sniffing before withdraw from the object in a new environment, while inhibition of SuM-DG circuit decreased the sniff latency but increased the time withdraw from the object (Fig. 3J and K).

Besides, activating or inhibiting SuM-DG circuit via optogenetic or chemogenetic approaches in an open field familiar to mouse but without alertness-related stimulation inputs did not alter average speed, freezing time, exploration rate, and percentage of time in center area of the open field, indicating there would be neither escape behavior nor increased anxiety level nor fear and stress when only manipulating SuM-DG circuit (Fig. S4).

Enhancing the SuM-DG circuit improves alertness and recognition

We further investigated the impact of SuM-DG circuit activity on recognition using chemogenetic and optogenetic approaches in WT mouse. The chemogenetic receptors either hM3Dq (Gq) or hM4Di (Gi), or halorhodopsin 3.0 (NpHR3.0) were expressed in SuM neurons projecting to DG (Fig. 4A). Through administration of clozapine-N-oxide (CNO) in mouse expressed Gq but not Gi, activation of SuM-DG circuit instead of inhibition it led to significant improvements in recognition, as evidenced by increased preference for novel objects in the NOR test (Fig. 4B). Interestingly, optogenetic inhibition SuM-DG circuit in looming stimulations attenuated the alertness-associated cognitive improvements in NOR (Fig. 4B). Together, these results suggest the necessary role of SuM-DG in learning and memory and alertness-associated cognitive improvements.

Fig. 4: Manipulating SuM-DG circuit alters alertness and recognition in ADHD.
figure 4

A Experimental timeline outlining the procedures for chemogenetic and optogenetic manipulations, including the schematic representation of viral vector delivery into the SuM. B Bar graph showing the percentage of time spent exploring a novel object in the NOR test under different experimental conditions. C Schematic of the experimental setup for visual looming stimulation, accompanied by a confocal microscopy image illustrating SuM illumination with blue light during optogenetic activation. D Bar graph comparing the average response times to 2# looming stimuli between control and ADHD rats under sham or optogenetic activation of SuM-DG circuit. E Line graph depicting the response times over five trials of looming stimulation of control or ADHD rats. F Bar graph showing the percentage of time spent exploring novel object in NOR test for control and ADHD rats.

Full size image

Then, to evaluate the therapeutic potential of targeting the SuM-DG circuit in ADHD, we examined the function of SuM-DG circuit via optogenetic manipulations in ADHD rats (Fig. 4C). Optogenetic activation of DG-projecting SuM neurons in ADHD rats resulted in a significant reduction in reaction times to the looming stimulus, indicating normalization of alertness levels (Fig. 4C–E). Additionally, this activation improved recognition in ADHD rats, demonstrated by increased time spent exploring novel objects in the NOR test (Fig. 4F). These findings suggest that hypoactivity of the SuM–DG circuit contributes to alertness deficits and cognitive impairments in SHRs, and that targeted activation can ameliorate these deficits.

By demonstrating that activation of the SuM-DG circuit can restore alertness and improve recognition in an ADHD animal model, we identify this pathway as a crucial neural substrate mediating the interplay between arousal regulation and cognitive function. The dense projections from SuM neurons to the DG provide a direct route through which arousal states influence hippocampal-dependent cognitive processes. The increased activation of this circuit during alertness-inducing stimuli, coupled with the ability to enhance alertness and cognition through optogenetic and chemogenetic modulation, underscores its potential as a therapeutic target in ADHD.

Alertness enhances recognition by facilitating long-term depression (LTD) in DG

Having established the critical role of the SuM–DG circuit in modulating alertness and recognition, we sought to elucidate the underlying synaptic mechanisms responsible for these enhancements. Given that synaptic plasticity within the DG is pivotal for processes, we hypothesized that plasticity might be a key mediator in the SuM–DG circuit’s facilitation of recognition.

To investigate this, we conducted in vivo electrophysiological recordings of field excitatory postsynaptic potentials (fEPSPs) in the DG of mice expressing either ChR2 or NpHR3.0 in SuM neurons (Fig. 5A). Following baseline recordings, we applied patterned light stimulation to SuM afferents projecting to the DG (Fig. 5A and B). Compared with optogenetic inhibition or sham stimulation, the optogenetic activation resulted in a significant and sustained decrease in fEPSPs amplitude, indicative of LTD induction. To confirm the involvement of LTD, we administered the specific LTD blocker Tat-GluR2A-3Y peptide prior to optogenetic stimulation. Pre-treatment with Tat-GluR2A-3Y effectively prevented the decrease in fEPSP amplitude caused by optogenetic activation of SuM (Fig. 5B). These results demonstrate that activation of SuM afferents facilitates LTD in the DG, and that this synaptic plasticity is essential for the observed electrophysiological changes.

Fig. 5: Alertness enhances recognition via facilitating LTD in SuM-DG circuit.
figure 5

A Schematic representation of viral vector delivery into DG neurons and the setup for in vivo electrophysiological recordings, illustrating the experimental approach to monitor synaptic activity influenced by SuM. B fEPSP recordings from DG neurons receiving projections from the SuM during light stimulation and pretreatment with Glu2A-3Y peptide, Bar 0.5 mV, 10 ms. C Above, experimental timeline depicting the procedures from habituation to testing, including periods of optogenetic activation, inhibition, and sham, as well as injections of the GluA2A-3Y peptide or a scrambled peptide control. D Bar graph showing the percentage of time spent exploring the novel object during the test phase under different experimental conditions.

Full size image

To determine whether LTD is required for the SuM-DG circuit’s enhancement of cognitive function, we conducted behavioral experiments using NOR test. Mice received intracerebral injections of 2 mg/kg Tat-GluR2A-3Y peptide or a scrambled control peptide 30 min before optogenetic manipulation (Fig. 5C). The SuM-DG projections were then activated or inhibited to the DG in mice expressing ChR2 or NpHR3.0, respectively. Behavioral assessments revealed that optogenetic activation of SuM fibers in the DG significantly increased the time spent exploring the novel object compared to sham controls (Fig. 5D), indicating enhanced recognition. Importantly, this improvement was abolished in mice pre-treated with the Tat-GluR2A-3Y peptide (Fig. 5D), demonstrating that LTD is necessary for the SuM-DG circuit-induced cognitive enhancements. Inhibition of SuM inputs to the DG did not produce significant changes in novel object preference, suggesting that basal activity of this pathway is not sufficient to modulate recognition in the absence of additional stimuli. These findings provide compelling evidence that alertness enhances recognition by facilitating LTD within the DG via the SuM-DG circuit.

Discussion

Our study uncovers a neural circuit mechanism that bridges alertness regulation and cognitive function deficits characteristic of ADHD. Utilizing SHR—an animal model that closely recapitulates the core symptoms of ADHD [27,28,29] —we demonstrated significant deficits in alertness and recognition. Through a combination of behavioral assays, c-Fos immunohistochemistry, optogenetics, chemogenetics, and electrophysiological recordings, we identified the supramammillary nucleus (SuM) as a critical neural substrate mediating alertness in response to both high-threat and mild stimuli. Moreover, we elucidated the pivotal role of the SuM–DG circuit in enhancing alertness and recognition via the facilitation of LTD in the DG.

These findings provide compelling evidence that hypoactivity within the SuM contributes to the alertness deficits observed in SHRs, and by extension, may underlie similar deficits in individuals with ADHD. The SuM has been implicated in a range of functions—including wakefulness, cognitive processing, locomotion, and motivation [16, 17, 25] —which aligns with the optimal stimulation theory of ADHD. This theory posits that individuals with ADHD exhibit chronic hypoarousal, leading them to seek additional physical activity and novel stimuli to achieve optimal arousal levels [8, 9]. Our observation that activation of the SuM–DG circuit enhances alertness and cognitive performance supports this model, suggesting that the SuM serves as a neural hub integrating arousal and motivational signals to modulate cognitive functions.

ADHD is frequently associated with sleep disturbances, including difficulties in initiating and maintaining sleep, as well as altered sleep architecture [33,34,35]. The SuM is known to play a crucial role in regulating sleep-wake cycles, contributing to the maintenance of wakefulness and transitions between sleep states [16]. Dysregulation of SuM activity may therefore contribute not only to alertness deficits but also to the sleep problems observed in ADHD. By influencing arousal states and sleep-wake regulation, the SuM may impact attentional capacities and cognitive performance. Our findings suggest that targeting the SuM could potentially ameliorate both sleep disturbances and cognitive deficits in ADHD, offering a more holistic therapeutic approach.

While arousal regulation has been a focus of ADHD research, the specific role of alertness—as a short-term, phasic arousal response to environmental stimuli [36] — has not been extensively investigated. Our study addresses this gap by demonstrating that alertness deficits can directly impair cognitive functions. The ability of the SuM–DG circuit to modulate alertness in response to both high-threat and mild stimuli highlights its importance in immediate attentional engagement and environmental responsiveness. By emphasizing the role of alertness, our work expands the understanding of arousal mechanisms in ADHD beyond tonic arousal levels, underscoring the need for further research into short-term arousal processes and their impact on cognitive function.

While our study focused on glutamatergic projections from the SuM to the DG, the SuM interacts with several neuromodulatory systems implicated in ADHD, including dopaminergic, noradrenergic, and serotonergic pathways [37]. The SuM’s connections with the ventral tegmental area and locus coeruleus suggest that it may influence, and be influenced by, dopamine and norepinephrine levels, which are critical for attention, motivation, and reward processing [38]. Dysregulation of these neurotransmitter systems is well-documented in ADHD and may interact with SuM activity to influence arousal, sleep-wake regulation, and cognitive functions [39,40,41]. Future studies should explore how modulation of the SuM–DG circuit affects these neuromodulatory systems, potentially offering a more comprehensive understanding of ADHD pathophysiology.

While SuM-DG research typically focuses on spatial coding and contextual novelty [17, 19], substantial evidence from multiple studies supports that the DG is essential for object recognition tasks and that disruption of DG function impairs performance on novel object recognition tasks [42,43,44,45]. Our data (Fig. S2A) showing selective DG activation following alertness-inducing stimuli supports the relevance of the SuM-DG pathway for object recognition, expanding our understanding of this circuit’s cognitive functions beyond spatial domains.

Our study advances understanding of neural mechanisms linking arousal regulation to cognitive function in ADHD. By identifying the SuM-DG circuit and LTD in the DG as key contributors to alertness and alertness-induced cognitive enhancements, we offer new perspectives on therapeutic strategies. These insights deepen our comprehension of ADHD’s neurobiological basis and highlight the importance of integrating molecular to circuit-level mechanisms when studying neuropsychiatric disorders. Future research should explore interactions with other systems, developmental factors, and the role of alertness to develop more effective and precise treatments for improving cognitive function in ADHD. Regarding translational potential, while direct modulation of the deep-located SuM presents technical challenges, non-invasive techniques targeting connected regions or behavioral interventions that enhance alertness could serve as effective alternatives based on our understanding of the SuM-DG circuit dynamics.