I am a neuroscientist with a background in bioengineering. I am interested in understanding how neuronal circuits in the brain give rise to sensory experience.
My research focuses on the use of behavioral measures and in vivo electrophysiology to understand how activity of neurons in the visual cortex correlates with perceptual behavior of animals. I use computational models to frame hypotheses about the underlying neuronal circuits that gave rise to experimentally observed responses of visual neurons. I make attempts to ground my hypotheses in our most current knowledge of neuroanatomy.
My current and future interests lie in leveraging our understanding of neuronal circuitry to manipulate brain activity through micro-stimulation (electric and/or pharmacological) across multiple cortical areas. This will help develop therapies for diseases and decipher the principles of brain function.
Double-click the blue markers [+] for further detail.[+] Tuning properties of cortical neurons
An important question in sensory neuroscience is, how do you perceive the world around you? What happens in the brain that enables you to see and recognize objects (stationary or moving) in your environment? In other words, how does perception of your environment arise from activity of neurons in your brain?
One way to answer this question is by studying selectivity. The visual system is selective for stimuli; it responds vigorously to some stimuli and less so to others. Neurons in the visual system are also selective. For e.g., neurons in middle temporal area (MT) are selective for direction of motion of the stimulus. Some neurons respond maximally only to stimulus moving at 45 degrees, while others respond only to stimulus moving at 90 degrees, so on and so forth. By measuring selectivity of the visual system of humans or animals (using behavioral methods) and comparing it to selectivity of neurons, one can start understanding the mechanisms of visual perception.
However, measuring selectivity of neurons is a non-trivial matter. Selectivity of neurons is encapsulated by the concept of classical receptive field. Classical receptive field is an important property of a neuron and summarizes the selectivity of the neuron to a wide range of stimulus properties. Some examples of these stimulus properties are:
- The physical location within an animal's field of vision (the receptive field center). A stimulus placed in the center strongly activates the neuron.
- The direction of motion of the stimulus
- Spatial and temporal frequency of the stimulus etc.
The classical receptive field of an MT neuron can be characterized by selectivity (or tuning) to the above mentioned stimulus properties. A stimulus placed in the classical receptive field center strongly activates the neuron by increasing its firing/spiking.
This notion of a classical receptive field has been challenged by the finding that when a stimulus is placed in the center as well as the region surrounding the center (the surround), the firing of the neuron changes as compared to when a stimulus was placed only in the center. Thus, the response of the neuron to a stimulus in the center of a receptive field can be changed by a stimulus placed in the surround. These interactions between the center and the surround are called contextual modulations. By changing the context (i.e. surrounding stimulus) one can change response of the neuron to the center.
So while the surround can modulate the center, the properties of the center are thought to be fairly stable. For a stimulus placed in the receptive field center, changing property of the stimulus along one dimension does not produce dramatic changes in tuning to another dimension. For e.g., changing the contrast (the difference between the lightest and darkest part of the stimulus) causes only modest changes in tuning of the neuron to spatial frequency.
For ease of measurement, most physiology studies evaluate tuning of neurons to a particular dimension (e.g. direction of motion) by measuring neuronal responses across different values of that dimensions, while keeping values of all other dimensions constant. For e.g., tuning of neurons to direction of motion is evaluated by measuring response of the neuron to all directions (0-360 degrees) while keeping contrast, spatial frequency and temporal frequency constant. Because of this, most studies are unable to capture the overall dynamism of the receptive field characteristics of neurons.
In our experiments, we measured receptive field properties across three stimulus dimensions; contrast, spatial frequency and temporal frequency. We did this by measuring the firing of neurons in area MT of alert behaving macaque monkeys.
We found that the tuning of cells in the middle temporal cortical area (MT) is more dynamic than it was previously believed. Neuronal tuning co-varied dramatically across contrast, spatial frequency and temporal frequency (in some cases as much as 30 fold). The effects of contrast and temporal frequency on changes in tuning were separable, indicating that some of the effects could be mediated by two separate circuit mechanisms. Since we used a fairly comprehensive set of stimulus parameters, we observed effects never seen previously in the literature. We thus showed that, contextual modulations exist not only in the receptive field surround, but also in the receptive field center.
We used a canonical inhibition-stabilized network model of interconnected excitatory and inhibitory cells to explain our experimental results. In such a model, each cell produces a neural wave when stimulated with a point stimulus (impulse response).
When we stimulated the circuit with a sinusoidal stimulus (similar to one used in the experiments), neural waves produced by individual cells caused constructive and destructive interference resulting in a pattern that looks similar to selectivity characteristics of cortical neurons.
When we changed contrast of the stimulus (again similar to our experimental design), tuning of the circuit to frequency of the stimulus changed substantially, in accord with our experimental results.
Our modeling results suggest that frequency tuning is an emergent property of the circuit. We hypothesize that tuning is not a property of individual neurons but a property the circuit (or circuits) to which the neuron belongs.
Such a shift of perspective from a cell-centered to a circuit-centered view has the potential of having a lasting impact on systems neuroscience.
One way to evaluate visual perceptual behavior of an observer is to measure contrast sensitivity. Contrast sensitivity, represented by the spatiotemporal contrast sensitivity function (CSF), is a large scale characteristic of visual performance. CSF measures the contrast at which a stimulus shown on a computer screen becomes just visible (contrast threshold) across a comprehensive set of visible stimuli.
Estimation of the CSF is important for basic vision research. Deficiencies in the CSF have been associated with several visual diseases and neuropathologies (Amblyopia, Parkinson’s disease, Multiple Sclerosis).
The issue with using the CSF as a standard clinical tool is that because of the sheer number of parameters to measure, estimation takes a lot of time. Thus due to practical reasons, a number of rapid methods of measurement of the CSF in humans have been developed.
It is beneficial to use similar methods of rapid measurement of CSF in non-human primates. Such methods would facilitate simultaneous measurements of behavior and neuronal responses which would then enable investigators to establish a direct relationship between neurophysiology and behavior
In this study, we developed and validated three methods for rapid measurements of CSF in two non-behaving primates. Measurements took a fraction of time compared to conventional slower methods.
The CSFs in both monkeys were similar to humans and each other. All three methods of measurement yielded similar results.
In this project, our goal was to develop novel light sensitive molecular structures as a therapy for blindness.
These molecular structure, when injected into the eye, would be capable of binding to post-synaptic receptors on retinal bipolar and ganglion cells, and after being stimulated by light, will be able to activate the receptors.
These structures consisted of the following components
- Anchor: antibody that can bind to receptors (e.g., a GABAc antibody)
- Chain: A flexible polyethylene glycol (PEG) chain
- Photoswitch: An azobenzene based photoswitch that changes configuration upon activation by light
- Effector: A receptor agonist or antagonist, that can activate receptors.
These structures would be injected into the eye and then specifically bind to receptors on bipolar and/or ganglion cells. Here they would stay in a dormant position until activated by light.
When light enters the eye and strikes the photoswitch on the structures, it changes the configuration of the structure. The PEG chain with the attached effector would then behave like a flexible robotic arm and activate the receptor.
Such molecular structures would then be able to restore sight in diseased retinas.
I tested components of the molecular structures, including the anchor and effector on GABAc receptor expressing neuroblastoma cells and GABAa expressing isolated retinal bipolar cells.
I developed a probe system for efficient delivery of components to GABAc receptors on neuroblastoma cells, while simultaneously recording from patch clamped cells. Such an approach was necessary to prevent wastage of molecular structures being developed by our collaborators.
Human neural stem cells were injected into stoke models of non-human primates. The stem cells were able to survive up to 105 days after an ischemic event and were able to partly undergo neuronal differentiation.
Description coming soon