Akhil Surapaneni Successfully Defended his PhD Thesis. Congratulations!!

On April 17, 2025 Akhil Surapaneni successfully defended his Phd thesis:

“Inhibition of Neural Activity in Brain Tumors: From Biology to Behavior”

Let’s Party!

We wish you all the best in your future career

 

Detailed Thesis Abstract:

A central tenet of neurosurgical oncology is achieving maximal safe resection of the tumor while preserving neurological function in the patient, a concept known as onco-functional balance. Our overarching hypothesis is that reorganizing functional circuits away from tumor-infiltrated brain regions can maximize surgical resection of the tumor while minimizing short-term functional deficits and potentially improving long-term functional outcomes.

We propose utilizing closed-loop neuromodulation and brain-computer interface (BCI) paradigms to decrease intra-tumoral functional representations in patients with glioma prior to surgery and transfer those functional circuits to peritumoral areas. In this thesis, we develop tools, and validate them in a series of pilot experiments in healthy volunteer population, that may be used for functional transfer in brain tumor patients.

First, we identified non-invasive stimulation methods, using repetitive transcranial magnetic stimulation (rTMS), for focal inhibition of endogenous activation patterns. We investigated the effect of low frequency rTMS in open-loop (n = 3 subjects) and a novel closed-loop paradigm (n = 3 subjects) on motor cortical activation patterns during continuous control of a motor imagery (MI) EEG brain-computer interface. Additionally, we tested continuous theta-burst stimulation (cTBS) on BCI control with a Riemannian geometry decoder (n = 5 subjects) that can correct for stimulation-induced non-stationarity. We found a trend for the effect of stimulation, across the three experiments, to reduce the posterior probability of decoding MI and increasing the posterior probability of decoding resting state. In the cTBS experiment, we found that EEG markers for activation showed a trend towards reduction in the electrode closest to the TMS following stimulation.
Next, we aim to develop a non-invasive BCI for providing feedback on cortical sources that can be used longitudinally and account for stimulation-induced non-stationarities, by combining traditional source inversion methods with a Riemannian geometry framework. The method was developed and validated using a dataset of 4 subjects. Additionally, 4 subjects controlled the source-space BCI in real-time, and pseudo-online analysis was performed to show the benefits of Riemannian geometry framework on source-level BCI control. Combining source-level BCI feedback and cTBS stimulation, we aimed to transfer sources of functional activation patterns from the hand motor area. The experiment was conducted over 8 days in 3 healthy volunteer subjects. If subjects successfully met the activation ratio criteria, functional electrical stimulation was delivered to contract the muscles whose movement the subject was imagining to control the BCI. Excessive activation inside the motor hotspot area was corrected with cTBS stimulation. As hypothesized, we found increased excitability and activation distal to the motor hotspot and reduced excitability and activation proximally to it due to training.
Lastly, we sought to understand how gliomas remodel normal neural tissue to gain insight in how to translate this into patients. We correlated the loss of inhibitory synaptic genes in glioma with overall survival using genomic analysis in a large patient database (n = 696 samples). In a separate model of patient-matched biopsies, we identified a preferential loss of parvalbumin-positive inhibitory interneurons in the samples that were previously characterized to have a greater degree of neuron-glioma interactions. Then, we identified a marker of interneuron loss in neural signaling, namely loss of gamma oscillations (30-80Hz), in tumor-infiltrated tissue as compared to normal regions of the brain using electrocorticography during a language naming tasks (n = 13 patients). However, phase-amplitude coupling of gamma oscillations was preserved.