Iftekhar Islam

Md Iftekhar Islam

Hi! I am a first-year Ph.D. student in the Medical Engineering Department at the University of South Florida. I’m working as a Research Assistant in the Advanced Biomedical Imaging lab at USF. My research is particularly centered around the groundbreaking fields of Photoacoustic and Ultrasound Tomography. Currently, I’m investigating the effectiveness of Low-Intensity Focused Ultrasound (LIFU) as a neuromodulation tool for modulating specific brain regions in mice to reduce their alcohol dependency behavior. The research involves identifying and targeting the neural circuits and brain areas associated with alcohol addiction. To achieve this aim, I’m studying the pathway and parameters of guiding the LIFU using Finite Element Modeling (FEM) based simulation.

Before joining USF, I worked as a Research Assistant at Research Hub.Inc, Bangladesh. In Research Hub, I developed different AI frameworks to process medical images and biological signals. I completed my Bachelor's in Electrical and Electronics Engineering from one of the reputed universities in Bangladesh. I have experience working as a Medical Physicist at the National Institute of Nuclear Medicine & Allied Science. Overall, I possess a strong background in Biomedical Engineering and have considerable experience working with multiple research groups and renowned organizations across the world.

Key research interests: Photoacoustic Tomography, Ultrasound Tomography, Image-Guided Therapy, Neuromodulation.

Email | Resume | Google Scholar | LinkedIn | Upwork

PhD Research Project

Title: A novel image-guided low intensity focused ultrasound (LIFU) based therapy for effectively treating alcohol use disorder (AUD)

The goal of this research is to develop a novel image-guided low-intensity focused ultrasound (LIFU) therapeutic modality to effectively treat alcohol use disorder (AUD). This project will evaluate behaviors related to affect and alcohol-seeking to understand the neurological impacts of LIFU, assess molecular brain responses post-treatment, and determine any adverse effects on neuronal function following chronic LIFU exposure.

The following sections describes different parts of the project that I already completed or currently working on.

1. Finite Element Method (FEM) based simulation to optimize LIFU parameters for targeting reward circuit

A finite element method (FEM)-based simulation model featuring a focused transducer and a detailed mouse head, including skull and brain tissue, was developed to investigate acoustic signal distribution within an animal brain. This simulation assessed several key parameters: 1) the efficiency of LIFU frequencies from 0.5 to 1.3 MHz, 2) the effects of sound velocity and attenuation in mouse brain tissue on LIFU propagation, and 3) how these parameters influence the power density and dimensions of the focal zone. After over 100 hours of simulation, aimed at replicating a realistic in-vivo experiment, findings indicated that skull sound attenuation markedly impacts LIFU power density, while brain tissue sound speed variations affect the focal zone’s size and location

Fig. 1. (a) Mouse head and focused transducer model, (b) FEM mesh in COMSOL, and (c) simulated acoustic field distribution.

Fig. 2. (a) Acoustic field distribution in the head with varied skull attenuation coefficients, and (b) acoustic signal intensity along the white dashed line shown in (a) through the center of the focus area.

Fig. 3. (a) Acoustic field distribution in the head with varied brain acoustic velocity, and (b) acoustic signal intensity along the white dashed line shown in (a) through the center of the focus area.

2. Photoacoustic Tomography (PAT) and Ultrasound Tomography (UT) combined image-guided treatment

An innovative image-guided technique that integrates photoacoustic tomography (PAT), ultrasound tomography (UT), and finite element (FE) modeling was proposed to optimize LIFU targeting in the mouse brain. The 2D PAT system will produce structural images of the mouse brain and will help to pinpoint the coordinates of the target area, while UT images will measure acoustic distribution, ensuring precise and optimized LIFU delivery.

Fig. 4. PAT system for LIFU guidance. A: Coronal slice of mouse brain around the point of stimulation B: PAT coronal slice showing the NAC. C: PAT coronal slice image showing the focus point at the NAc. D: Our PAT system used for LIFU guidance. E: Coronal PAT Image and identified structural landmarks: Bregma and lambdoid

Initially, the nucleus accumbens (NAc), ventral tegmental area (VTA), and medial prefrontal cortex (mPFC) will be ultrasonically stimulated in vivo. These regions are targeted because they house neurons critical to alcohol-seeking behaviors. Crossed High Alcohol Preferring (cHAP) mice, known for their inherent addiction to alcohol, will be used in this therapeutic study.

3. Validation of the UT and PAT imaging platforms with phantom experiments

Several phantom experiments using two distinct phantom types, one for UT and another for PAT, were conducted to validate the quantitative capabilities of our UT and PAT imaging systems. For the UT phantom, glycerin was used to emulate the sound speed of brain tissue, adjustable from 1500 to 1750 m/s by varying the glycerin concentration, and graphite to mimic tissue sound attenuation, adjustable from 0.3 to 1 dB/cm with varying graphite concentration. I cross-validated the acoustic parameters of these phantoms by measuring their physical properties with a 1 MHz unfocused transducer and a needle hydrophone. This involved comparing the literature values and physical acoustic values of the phantoms with the average values from our UT system’s quantitative images, achieving an error margin of less than 6%.

Fig. 5. (a) Schematic of phantom with an internal target. (b) Table depicting the parameters of four phantoms with varying sizes and acoustic properties. (c) Photographs of the four phantoms.

Fig. 6. Reconstructed UT images of acoustic velocity and attenuation.

Additionally, I took 2D PAT images of various layers of the mouse brain to study their structural distribution and to validate the system's precision.

Fig. 7. 2D PAT mouse brain image

Some other research works

	3D Prototype Designing for Research Work As a part of my research work, I often design 3D prototypes and tools. I mainly use the SOLIDWORKS software for 3D designing purposes. I designed a frame that was able to move a 12-inch by 10-inch Ultrasound Transducer (about 10 lb.) in multiple directions as well as uplift it vertically. The primary purpose of the frame was to hold the transducer tightly and move it in multiple directions so that the whole-body imaging of a mouse could be done without any hassle. The frame was designed in 12 individual parts and then printed out using a 3D printer.
	Deep Learning in Lung and Colon Cancer Classifications Krishna Mridha, MD. Iftekhar Islam, Shamin Ashfaq, Masrur Ahsan Priyok, Dipayan Barua. Paper Link: IEEE Explore, An AI-based framework to classify and differentiate five types of colon and lung cancer cells. The framework uses Digital Image processing (DIP) and Deep Learning (DL) algorithms to process and evaluate histopathological images of cancer cells.
	DBNet: Detect Diabetic Retinopathy to Stop Blindness Before It’s Too Late Krishna Mridha, Meghla Monir Shorna, Nazmul Arefin, Ananya Ritu, MD Minhazul Alam Chowdhury, MD. Iftekhar Islam Paper Link: IEEE Xplore A hybrid autoencoder framework that combines Convolutional Neural Networks (CNN) and Long-Short Term Memory (LSTM) to enhance the accuracy of adaptive retinopathy prediction from fundus images of the retina.
	ML-DP: A Smart Emotion Detection System for Disabled Person to Develop a Smart City Krishna Mridha, MD. Iftekhar Islam, Meghla Monir Shorna, Masrur Ahsan Priyok. Paper Link: IEEE Xplore A smart emotion recognition system that uses a convolutional neural network to analyze the facial landmarks of impaired people to comprehend their emotional state. The system is also able to inform the caregiver in case of a sudden breakdown of emotions of the impaired individual.
	An indoor navigation system for visually impaired people using a pathfinding algorithm and a wearable cap MD. Iftekhar Islam, MD. Forhadur Rahman, MD. Maruf Hossain Raj, Sabbir Hossen, Shantanu Nath, Dr. Mohammad Hasan Imam Paper Link: IEEE Xplore The indoor-based navigation system, specially designed for visually impaired people, consists of two modules: a cap designed with IR receivers and sensors, and the architecture of the area where the navigation system works by guiding the user.
	Bioactivity classification of SARS-CoV-2 Proteinase using Machine Learning Approaches Fatema Begum, Krishna Mridha, Md Golam Rabbani, Shamin Ashfaq, MD. Iftekhar Islam, Sapna Sinha Paper Link: IEEE Xplore A novel classifier model that forecasts Bioactivity of SARS-CoV-2 Proteinase more accurately than other existing models. The model uses a variety of classification techniques, including the SVM, Random Forest, LR, KNN, and Naive Bayes. The major goal of applying the parametric technique is to improve the performance of a well-known SARS-CoV-2 single proteins dataset derived from the well-known “ChEMBL” database.

Professional Experience

PhD Student, University of South Florida (August 2023 - Present)
IELTS Instructor, London Educators (April 2023 – August 2023)
Research Assistant, Research Hub Inc. (January 2022 - August 2023)
Medical Physicist, NINMAS (January 2020 – November 2020)

Based on Jon Barron's site.
Last updated December 2021.