Arvind Kumar

This paper investigates reaction and recovery times of leaky integrate-and-fire neurons in response to changes in mean input drive or input fluctuation level.

The paper shows that spike-timing-dependent plasticity (STDP) selects connections with delays matching network oscillation periods, optimizing signal transmission efficiency.

The paper investigates how inhibitory gain and hub architecture contribute to neural circuit resilience against degeneration by analyzing spiking networks with controlled pruning strategies.

This study identifies high-impact locations in neural networks that can control sequential activity through local synaptic changes.

This study shows that NMDA receptors adjust neuronal gain across the cortical hierarchy, allowing integration of dense presynaptic inputs.

This study shows that the spectral slope of population activity cannot reliably predict excitation-inhibition balance in neural networks.

The paper shows that inhibitory neurons with fewer recurrent connections can reduce synchrony caused by over-represented excitatory motifs in neuronal networks.

This study shows how weak neural sequences can overcome competition by increasing connectivity or cooperating with other sequences, affecting their speed and reactivation.

The paper constructs a neural circuit model of mesocorticolimbic dopamine signaling to explain schizophrenia's symptoms through disrupted reward learning and value representation.

This study investigates how fast-spiking interneurons (FSIs) in the striatum modulate input correlation and trial-to-trial variability, revealing their role in both exploration and stability during learning.

This study shows that RNNs can learn task structure with random feedback and biological constraints without needing offline updates or eligibility traces.

This paper investigates how neural trajectories and their subspaces are dynamically controlled in the central nervous system through mechanisms like gain modulation and clustered inhibition.

The paper shows that correlated asymmetric connectivity in neural networks generates spatiotemporal activity sequences, while symmetric connectivity produces stationary spatial bumps.

This in-silico study predicts how Parkinson's disease alters striatal network connectivity and glutamatergic drive to striatal projection neurons.

The paper shows that spatially heterogeneous inhibition allows sequential neural activity to project onto task-specific subspaces, enabling fast switching controlled by top-down signals.

The paper demonstrates that the mouse superior colliculus generates center-surround visual computations through local circuits, independent of top-down input.

The paper shows that input variance ratios and covariances determine output trial-by-trial variability in a cortical microcircuit model with different interneuron subtypes.

The paper investigates how excitation and inhibition delays in cerebellar Purkinje cells modulate output through a feedforward inhibitory pathway in mice.

This study analyzes resting-state MEG data to show that Parkinson's disease is associated with widespread changes in cortical aperiodic activity, specifically increased λ values across most brain regions.

The paper shows that decoding time and stimulus dimensionality affect the optimal shape of tuning curves, favoring single-peaked over periodic in early visual areas due to a speed-accuracy trade-off.

This paper proposes a model where vector-valued dopamine signals enhance learning of continuous outputs in the striatum using a recurrent network and a local plasticity rule.

The paper models the visual microcircuit in the mouse primary visual cortex, showing that superficial and deep layers implement bistable switches based on inhibitory connectivity between SST, PV, and VIP cells, leading to state transitions with distinct oscillatory patterns.

This study uses a network model of the basal ganglia to show that low-dopamine state alters transient responses and oscillatory activity through changes in direct pathway and recurrent connectivity within the globus pallidus externa.

The paper analyzes the structure and dynamics of biological neuronal networks to determine how network topology influences activity, finding that effective synaptic weight, not topology class, is the key predictor of spiking dynamics.

The paper shows how short-term synaptic plasticity helps neurons distinguish sparse signals from background activity in a presynaptic population.

This study extends an EI neuron network model to include an E population with transmission delays, showing how critical coupling strength and delay modulate oscillatory and stationary states.

This paper shows that adding feedback connections to modular neuronal networks improves communication through resonance by enabling faster and more reliable propagation of single spike volleys.

This study uses numerical simulations to show that STN firing rates drive beta oscillations, while GPe firing rates and spike bursting have state-dependent effects on beta band oscillations in the STN-GPe network.

The paper develops a neurochemical model showing D1 and D2 dopamine receptors respond similarly to tonic and phasic signals due to receptor kinetics and abundance.

This paper shows that anisotropic and correlated connectivity in neuronal networks can generate spatio-temporal activity sequences without requiring external training or manual embedding.