FKFumi Katsuki

Fumi Katsuki

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Fumi Katsuki has a total of 31 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Early Loss of Deep Restorative Sleep and Auditory Stimulus Evoked 40-Hz activity of Hippocampal Parvalbumin Neurons in the APP/PS1 Mouse Model of Alzheimer’s Disease

    by Fumi Katsuki, James M. McNally, Dmitry Gerashchenko, David S. Uygun, Alyssa Tyler, John G. McCoy, James T. McKenna, Ritchie E. Brown

    Sleep abnormalities and dysfunction of gamma band (30–80 Hz) activity generated by parvalbumin (PV) interneurons are early characteristics of Alzheimer’s disease (AD) which correlate with the severity of amyloid-β deposition (Aβ) and cognitive impairment. However, the timing of these alterations in vivo with respect to disease progression is unclear. Here, in longitudinal recordings from APP/PS1/PV-cre (AD mice) from 3–6 months, we found reduced sleep slow-wave power (0.5–4 Hz) in hippocampus and medial prefrontal cortex in AD mice as young as 3 months old, compared to non-AD (PV-cre) mice, well before overt pathology. This finding was primarily due to reductions in the NREM delta range (1.5–4 Hz), a hallmark of restorative functions of sleep. In contrast, beta (15–30 Hz) power linked to insomnia was significantly higher across all sleep-wake states. Loss of deep NREM sleep was not compensated by an increase in NREM sleep time, instead NREM sleep during the dark (active) phase was slightly but significantly lower in AD mice. 40-Hz auditory steady-state responses and associated evoked calcium responses of hippocampal PV neurons recorded using fiber photometry were also impaired by 3 months old. However, Y-maze performance in 3- and 6-month-old AD mice was not significantly different from non-AD mice. These results reveal reduced deep sleep and PV-associated 40-Hz activity as very early changes amenable to early intervention occurring prior to cognitive deficits. Furthermore, they establish APP/PS1 mice as a good model to causally test the relationship between sleep, PV neuronal activity and amyloid-mediated pathology.

  • Open Access

    Synaptic GABA dysfunction of thalamocortical neurons impairs sleep spindle morphology and recovery from fearful memories

    by Fumi Katsuki, Meghan C Bauer, Mitchell J Vaughn, Victoria A. Lombardi, Ritchie E. Brown, Julie S Haas, Radhika Basheer, David S. Uygun

    Sleep spindles are rhythmic electroencephalographic signatures of non-rapid-eye-movement sleep. Their dysregulation has been implicated in several neuropsychiatric illnesses. Spindles have a characteristic waxing and waning shape, but the cellular and circuit mechanisms controlling their shape are not well understood. Recent but sparse research has implied that sleep spindle shape becomes abnormal in post-traumatic stress disorder (PTSD). PTSD patients have dysfunctional GABA A receptors in midline thalamic regions, areas involved in the orchestration of sleep spindles. We modelled this GABA A dysfunction within thalamocortical (TC) neurons using localized CRISPR-Cas9 technology to test the hypothesis that GABA dysfunction would dysregulate sleep spindle shape and cause symptoms of PTSD, in mouse model behavioral evaluations. We found sleep spindles were shorter and abnormally shaped, having lost their characteristic waxing and waning shape, in mice with GABA A receptor knock-down in TC neurons (TC-α1KD). TC-α1KD mice failed to recover from learned fearful reactions following an aversive stimulus. We tested this with a contextual fear conditioning paradigm using electric foot shocks. A control group with intact GABA A receptors successfully habituated to the fear conditioned location in subsequent visits to that context without foot shocks. In contrast, TC-α1KD mice never habituated, suggesting abnormally extended fearful memories. The number of inhibitory post synaptic currents in TC neurons were significantly decreased in vitro , confirming an effective knock-down. Our results imply that abnormally shaped sleep spindles may serve as a biomarker of GABA A receptor dysfunction in TC neurons which may be involved in abnormal fear processing in PTSD. We postulate GABA A receptor dysfunction in TC neurons may be underlying pathophysiology of PTSD and our findings here may inspire the development of screens, diagnostics and objective characteristics of stress related disorders, including PTSD.

  • Open Access

    Neuronal PAS domain 1 identifies a major subpopulation of wakefulness-promoting GABAergic neurons in the basal forebrain

    by Timothy A. Troppoli, Chun Yang, Fumi Katsuki, David S. Uygun, Ilyan Lin, David D. Aguilar, Tristan Spratt, Radhika Basheer, James M. McNally, C. Savio Chan, et al.

    Here, we describe a group of basal forebrain (BF) neurons expressing neuronal Per-Arnt-Sim (PAS) domain 1 (Npas1), a developmental transcription factor linked to neuropsychiatric disorders. Immunohistochemical staining in Npas1-cre-2A-TdTomato mice revealed BF Npas1 + neurons are distinct from well-studied parvalbumin or cholinergic neurons. Npas1 staining in GAD67-GFP knock-in mice confirmed that the vast majority of Npas1 + neurons are GABAergic, with minimal colocalization with glutamatergic neurons in vGlut1-cre-tdTomato or vGlut2-cre-tdTomato mice. The density of Npas1 + neurons was high, five to six times that of neighboring cholinergic, parvalbumin, or glutamatergic neurons. Anterograde tracing identified prominent projections of BF Npas1 + neurons to brain regions involved in sleep–wake control, motivated behaviors, and olfaction such as the lateral hypothalamus, lateral habenula, nucleus accumbens shell, ventral tegmental area, and olfactory bulb. Chemogenetic activation of BF Npas1 + neurons in the light period increased the amount of wakefulness and the latency to sleep for 2 to 3 h, due to an increase in long wake bouts and short NREM sleep bouts. NREM slow-wave and sigma power, as well as sleep spindle density, amplitude, and duration, were reduced, reminiscent of findings in several neuropsychiatric disorders. Together with previous findings implicating BF Npas1 + neurons in stress responsiveness, the anatomical projections of BF Npas1 + neurons and the effect of activating them suggest a possible role for BF Npas1 + neurons in motivationally driven wakefulness and stress-induced insomnia. Identification of this major subpopulation of BF GABAergic neurons will facilitate studies of their role in sleep disorders, dementia, and other neuropsychiatric conditions involving BF.


  • Open Access

    Sleep-Deep-Learner is taught sleep–wake scoring by the end-user to complete each record in their style

    by Fumi Katsuki, Tristan J Spratt, Ritchie E Brown, Radhika Basheer, David S Uygun

    Sleep–wake scoring is a time-consuming, tedious but essential component of clinical and preclinical sleep research. Sleep scoring is even more laborious and challenging in rodents due to the smaller EEG amplitude differences between states and the rapid state transitions which necessitate scoring in shorter epochs. Although many automated rodent sleep scoring methods exist, they do not perform as well when scoring new datasets, especially those which involve changes in the EEG/EMG profile. Thus, manual scoring by expert scorers remains the gold standard. Here we take a different approach to this problem by using a neural network to accelerate the scoring of expert scorers. Sleep-Deep-Learner creates a bespoke deep convolution neural network model for individual electroencephalographic or local-field-potential (LFP) records via transfer learning of GoogLeNet, by learning from a small subset of manual scores of each EEG/LFP record as provided by the end-user. Sleep-Deep-Learner then automates scoring of the remainder of the EEG/LFP record. A novel REM sleep scoring correction procedure further enhanced accuracy. Sleep-Deep-Learner reliably scores EEG and LFP data and retains sleep–wake architecture in wild-type mice, in sleep induced by the hypnotic zolpidem, in a mouse model of Alzheimer’s disease and in a genetic knock-down study, when compared to manual scoring. Sleep-Deep-Learner reduced manual scoring time to 1/12. Since Sleep-Deep-Learner uses transfer learning on each independent recording, it is not biased by previously scored existing datasets. Thus, we find Sleep-Deep-Learner performs well when used on signals altered by a drug, disease model, or genetic modification.


  • Open Access

    Sleep-Deep-Net learns sleep wake scoring from the end-user and completes each record in their style

    by Fumi Katsuki, Tristan J. Spratt, Ritchie E. Brown, Radhika Basheer, David S. Uygun

    Sleep-wake scoring is a time-consuming, tedious but essential component of clinical and pre-clinical sleep research. Sleep scoring is even more laborious and challenging in rodents due to the smaller EEG amplitude differences between states and the rapid state transitions which necessitate scoring in shorter epochs. Although many automated rodent sleep scoring methods exist, they do not perform as well when scoring new data sets, especially those which involve changes in the EEG/EMG profile. Thus, manual scoring by expert scorers remains the gold-standard. Here we take a different approach to this problem by using a neural network to accelerate the scoring of expert scorers. Sleep-Deep-Net (SDN) creates a bespoke deep convolution neural network model for individual electroencephalographic or local-field-potential records via transfer learning of GoogleNet, by learning from a small subset of manual scores of each EEG/LFP record as provided by the end-user. SDN then automates scoring of the remainder of the EEG/LFP record. A novel REM scoring correction procedure further enhanced accuracy. SDN reliably scores EEG and LFP data and retains sleep-wake architecture in wild-type mice, in sleep induced by the hypnotic zolpidem, in a mouse model of Alzheimer’s disease and in a genetic knock-down study, when compared to manual scoring. SDN reduced manual scoring time to 1/12. Since SDN uses transfer learning on each independent recording, it is not biased by previously scored existing data sets. Thus, we find SDN performs well when used on signals altered by a drug, disease model or genetic modification. STATEMENT OF SIGNIFICANCE: Sleep medicine is often critically advanced by translational research based on in vivo electrophysiologic mouse data. A necessary but time-consuming step in this field is scoring epochs of recordings into wakefulness, non-rapid-eye-movement sleep and non-rapid-eye-movement sleep. Despite efforts to automate this, manual scoring remains the gold-standard since automatic methods poorly handle data that is not similar enough to data used during development. Here, we describe a novel automated sleep scoring method that involves retraining a deep-convolution-neural-net capable of computer vision to score sleep-wake patterns after learning from a small set of manual scores within a record. This avoids biasing the model to expect data to be the same as its training set from previous records.


  • Open Access

    Neuronal PAS domain 1 identifies a major subpopulation of wakefulness-promoting GABAergic neurons in basal forebrain

    by Timothy A. Troppoli, Chun Yang, Fumi Katsuki, David S. Uygun, Ilyan Lin, David D. Aguilar, Tristan Spratt, Radhika Basheer, James M. McNally, C. Savio Chan, et al.

    Here we describe a novel group of basal forebrain (BF) neurons expressing neuronal PAS domain 1 (Npas1), a developmental transcription factor linked to neuropsychiatric disorders. Immunohistochemical staining in Npas1-cre-2A-TdTomato mice revealed BF Npas1 + neurons are distinct from well-studied parvalbumin or cholinergic neurons. Npas1 staining in GAD67-GFP knock-in mice confirmed that the vast majority of Npas1 + neurons are GABAergic, with minimal colocalization with glutamatergic neurons in vGlut1-cre-tdTomato or vGlut2-cre-tdTomato mice. The density of Npas1 + neurons was high, 5-6 times that of neighboring cholinergic, parvalbumin or glutamatergic neurons. Anterograde tracing identified prominent projections of BF Npas1 + neurons to brain regions involved in sleep-wake control, motivated behaviors and olfaction such as the lateral hypothalamus, lateral habenula, nucleus accumbens shell, ventral tegmental area and olfactory bulb. Chemogenetic activation of BF Npas1 + neurons in the light (inactive) period increased the amount of wakefulness and the latency to sleep for 2-3 hr, due to an increase in long wake bouts and short NREM sleep bouts. Non-REM slow-wave (0-1.5 Hz) and sigma (9-15 Hz) power, as well as sleep spindle density, amplitude and duration, were reduced, reminiscent of findings in several neuropsychiatric disorders. Together with previous findings implicating BF Npas1 + neurons in stress responsiveness, the anatomical projections of BF Npas1 + neurons and the effect of activating them suggest a possible role for BF Npas1 + neurons in motivationally-driven wakefulness and stress-induced insomnia. Identification of this major subpopulation of BF GABAergic neurons will facilitate studies of their role in sleep disorders, dementia and other neuropsychiatric conditions involving BF. SIGNIFICANCE STATEMENT: We characterize a group of basal forebrain (BF) neurons in the mouse expressing neuronal PAS domain 1 (Npas1), a developmental transcription factor linked to neuropsychiatric disorders. BF Npas1 + neurons are a major subset of GABAergic neurons distinct and more numerous than cholinergic, parvalbumin or glutamate neurons. BF Npas1 + neurons target brain areas involved in arousal, motivation and olfaction. Activation of BF Npas1 + neurons in the light (inactive) period increased wakefulness and the latency to sleep due to increased long wake bouts. Non-REM sleep slow waves and spindles were reduced reminiscent of findings in several neuropsychiatric disorders. Identification of this major subpopulation of BF GABAergic wake-promoting neurons will allow studies of their role in insomnia, dementia and other conditions involving BF.

  • Open Access

    Knockdown of GABAA alpha3 subunits on thalamic reticular neurons enhances deep sleep in mice

    by David S. Uygun, Chun Yang, Elena R. Tilli, Fumi Katsuki, Erik L. Hodges, James T. McKenna, James M. McNally, Ritchie E. Brown, Radhika Basheer

    Identification of mechanisms which increase deep sleep could lead to novel treatments which promote the restorative effects of sleep. Here, we show that knockdown of the α3 GABAA-receptor subunit from parvalbumin neurons in the thalamic reticular nucleus using CRISPR-Cas9 gene editing increased the thalamocortical delta (1.5–4 Hz) oscillations which are implicated in many health-promoting effects of sleep. Inhibitory synaptic currents in thalamic reticular parvalbumin neurons were strongly reduced in vitro. Further analysis revealed that delta power in long NREM bouts prior to NREM-REM transitions was preferentially affected by deletion of α3 subunits. Our results identify a role for GABAA receptors on thalamic reticular nucleus neurons and suggest antagonism of α3 subunits as a strategy to enhance delta activity during sleep.


  • Characterization of basal forebrain glutamate neurons suggests a role in control of arousal and avoidance behavior

    by James T. McKenna, Chun Yang, Thomas Bellio, Marissa B. Anderson-Chernishof, Mackenzie C. Gamble, Abigail Hulverson, John G. McCoy, Stuart Winston, Erik Hodges, Fumi Katsuki, et al.

    The basal forebrain (BF) is involved in arousal, attention, and reward processing but the role of individual BF neuronal subtypes is still being uncovered. Glutamatergic neurons are the least well-understood of the three main BF neurotransmitter phenotypes. Here we analyzed the distribution, size, calcium-binding protein content and projections of the major group of BF glutamatergic neurons expressing the vesicular glutamate transporter subtype 2 (vGluT2) and tested the functional effect of activating them. Mice expressing Cre recombinase under the control of the vGluT2 promoter were crossed with a reporter strain expressing the red fluorescent protein, tdTomato, to generate vGluT2-cre-tdTomato mice. Immunohistochemical staining for choline acetyltransferase and a cross with mice expressing green fluorescent protein selectively in GABAergic neurons confirmed that cholinergic, GABAergic and vGluT2+ neurons represent distinct BF subpopulations. Subsets of BF vGluT2+ neurons expressed the calcium-binding proteins calbindin or calretinin, suggesting that multiple subtypes of BF vGluT2+ neurons exist. Anterograde tracing using adeno-associated viral vectors expressing channelrhodopsin2-enhanced yellow fluorescent fusion proteins revealed major projections of BF vGluT2+ neurons to neighboring BF cholinergic and parvalbumin neurons, as well as to extra-BF areas involved in the control of arousal or aversive/rewarding behavior such as the lateral habenula and ventral tegmental area. Optogenetic activation of BF vGluT2+ neurons elicited a striking avoidance of the area where stimulation was given, whereas stimulation of BF parvalbumin or cholinergic neurons did not. Together with previous optogenetic findings suggesting an arousal-promoting role, our findings suggest that BF vGluT2 neurons play a dual role in promoting wakefulness and avoidance behavior.


  • Open Access

    CRISPR knockdown of GABA A alpha3 subunits on thalamic reticular neurons enhances deep sleep

    by David S. Uygun, Chun Yang, Elena R. Tilli, Fumi Katsuki, James T. McKenna, James M. McNally, Ritchie E. Brown, Radhika Basheer

    Identification of mechanisms which increase deep sleep could lead to novel treatments which promote the restorative effects of sleep. Here, knockdown of the α3 GABA A -receptor subunit from parvalbumin neurons in the thalamic reticular nucleus using CRISPR-Cas9 gene editing increased the thalamocortical delta oscillations implicated in many health-promoting effects of sleep. Inhibitory synaptic currents were strongly reduced in vitro . Effects were most pronounced in the mouse sleep (light) period. Further analysis identified a novel deep-sleep state in mice prior to NREM-REM transitions which was preferentially affected by deletion of α3 subunits. Our results identify a functional role for GABA A receptors on Thalamic Reticular Nucleus neurons and suggest antagonism of α3 subunits as a strategy to enhance deep sleep. One Sentence Summary: Selective genetic knockdown of the major α subunit of GABA A receptors present in the thalamic reticular nucleus enhanced deep sleep in mice.