<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Wennie Xu on Medium]]></title>
        <description><![CDATA[Stories by Wennie Xu on Medium]]></description>
        <link>https://medium.com/@xxxxu0714?source=rss-b35b8c371f5a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*iIBvUHqXM6i9wzBJvfb7gg.jpeg</url>
            <title>Stories by Wennie Xu on Medium</title>
            <link>https://medium.com/@xxxxu0714?source=rss-b35b8c371f5a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 19:53:02 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@xxxxu0714/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Towards Early Detection: Classifying Alzheimer’s EEG Signals with Machine Learning]]></title>
            <link>https://medium.com/@xxxxu0714/towards-early-detection-classifying-alzheimers-eeg-signals-with-machine-learning-885a1f1464c8?source=rss-b35b8c371f5a------2</link>
            <guid isPermaLink="false">https://medium.com/p/885a1f1464c8</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[brain]]></category>
            <category><![CDATA[neuroscience]]></category>
            <category><![CDATA[brain-computer-interface]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Wennie Xu]]></dc:creator>
            <pubDate>Sun, 30 Nov 2025 06:55:02 GMT</pubDate>
            <atom:updated>2025-11-30T06:55:02.210Z</atom:updated>
            <content:encoded><![CDATA[<p>Alzheimer’s disease is notoriously difficult to diagnose early, especially before noticeable memory loss begins. However, recent research suggests that subtle patterns in brainwave activity - especially in the alpha, theta, delta, and gamma frequency bands - may reveal early neural decline long before symptoms appear.</p><p>Using machine learning, it may be possible to detect these patterns automatically.</p><p>This project explores that possibility by using EEG recordings from healthy individuals and individuals with Alzheimer’s, extracting meaningful signal features, and training an AI model to distinguish between the two.</p><h4><strong>The science behind this:</strong></h4><p>This project aims to take advantage of the changes in brain waves after a patient is developing or has already developed Alzheimer’s disease.</p><p>The healthy brain contains tens of billions of neurons, cells that process and transmit information via electrical and chemical signals. These cells send messages between each other, as well as from the brain to the muscles to initiate movement. Simply speaking, Alzheimer’s disease would disrupt or hinder this communication, causing cognitive decline as neurons eventually lose their function and die.</p><p>At first, the connections between parts involved in memory (entohinal cortex and hippocampus) get disrupted. Later, the cerebral cortex would get affected (thought, memory language). Eventually, many other areas of the brain are damaged.</p><p>This is the result of abnormal protein build-up in the brain known as tau-tangles and amyloid plaques.</p><p>Tau proteins, in healthy cells, are actually useful for the cell’s function. However, in Alzheimer’s disease, a cascade of events trigger to be hyper-phosphorylated (in other words, have excessive number of phosphate (PO4) added onto it. Hence, they begin clumping together to form the so called “tangles”, which collaspes the internal transport system of the neuron, preventing it from function and eventually leading to death.</p><p>Amyloid plaques, on the other hand, are abnormal clumps of the beta-amyloid proteins that build up int he spaces between nerve cells, formed when a larger protein (amyloid precursor protein) is broken down incorrectly. These plaques, when formed in the brain, results in the dying off of neurons, leading to cognitive decline.</p><p>Currently, the diagnostic process of Alzheimer’s involves cognitive and medical tests to assess memory and brain functions. However, many issues exist with the reliability and accuracy of biomarkers, especially in early stages, which results in the disease being discovered a lot later than it should. This would result in a delay of treatment, hence resulting in a lowered effectiveness of current drugs.</p><p>However, one important aspect of Alzheimers disease is that it disrupts brain wave patterns: typically causing a decrease of faster brain wave oscillations (particularly in gamma rhythms, &gt;30Hz) and an increase in slower waves like delta and theta waves (1–4 Hz, and 4–8 Hz, respectively).</p><p>These changes in rhythms would usually come before any apparent symptoms of Alzheimer’s disease. Hence, an early detection system would be especially useful in the detection of Alzheimer’s disease, which would increase the efficiency and usefulness of treatments.</p><h4>The project</h4><p>In this project, I built a simplified version of such a detector using Machine Learning. Even though much training is still needed to improve the true accuracy of this, it offers a simplified gist of something comprehensive and important in the future for detecting Alzheimer’s disease.</p><p>This is an extension of my last project, a simplified version of a brain computer interface where I decoded brain signals for the brain being awake vs at rest and visualized it into graphs.</p><p>To start, EEG (Electroencephalogram, a type of brain computer interface where electrodes are placed on the scalp) data is collected from two groups, those diagnosed with Alzheimer’s vs Healthy individuals. Then, the data was stored into files, organized carefully into seperate folders for healthy and Alzheimer’s participants. This helps the computers learn specific patterns to each group.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4Vej-q_gw0Fet1kp7DR8OA.png" /><figcaption>Organized data to train the model with</figcaption></figure><p>Before continuing on, I decided to first visualize the data to make sure it is present</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1022/1*RLAzdULU39wh0Oqq4woXiQ.png" /><figcaption>Visual representation of EEG recording from an Alzheimer’s patient. Left hand column markers meaning (Letters in front: Part of the brain electrode is placed on. F=Frontal Lobe; C= Central sulcus; P= Parietal, T=Temporal, P= Parietal. Numbers after: showcases position on the head. Odd numbers (1, 3, 5, 7): left hemisphere; even numbers: (2, 4, 6, 8): right hemisphere and z: centre of head)</figcaption></figure><p>Next, I developed a method to extract important features from the EEG data. Features are measurable aspects of the brain waves, like the average power in the alpha or gamma frequency bands, or how often bursts of these waves appear. By calculating these features for each file, we can summarize complex brain activity into numbers that a computer can understand.</p><p>The visualization for EEG data looked something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ujTvJ1_gq1zfVRR2Z7kDgQ.png" /><figcaption>Violin plot of healthy vs Alzheimer’s brain waves</figcaption></figure><p>The data was visualized using both Violin plots, shown above.</p><p>These plots measure out the distribution of certain biomarkers between healthy patients and those with Alzheimer’s disease. For example, the average strength of alpha waves (alpha_mean), the ratio between power of theta waves and alpha waves (theta_alpha_ratio), and the counts of alpha and gamma bursts: brief spikes of alpha and gamma activity.</p><p>The parts where the graphs are wider indicates the areas where the most data points are concentrated, whereas the dotted lines represent medians and quartiles in the data.</p><p>Once I had these features visualized, we built a dataset where each participant’s brain wave features were paired with their label:“healthy” or “Alzheimer’s.” Now, the model could be trained to recognize the patterns associated with Alzheimer’s. This trained model can then be used the to predict whether new EEG recordings belong to a healthy brain or one at risk for Alzheimer’s.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tDcnVsrbt35Zek8ZoHFgEg.png" /><figcaption>Training the model. Notice how the accuracy below is 1.00 (100% accuracy). This would mean that the model would need further training and larger datasets in order to accurately assess its true accuracy.</figcaption></figure><p>Finally, we also set up a way to test new data. If someone uploads an EEG recording, the system can process it in the same way, extract the same features, and then use the trained model to make a prediction. This completes the loop from collecting brain wave data to providing meaningful predictions that could help in early detection of Alzheimer’s disease.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RHd1DXUsmIiKL9fDynkgoA.png" /><figcaption>Example of the uploading files process. The model predicts Alzheimer’s brain waves here.</figcaption></figure><h4>Conclusion:</h4><p>Overall, this process was a very fun project that I enjoyed doing. Visualizing the data through violin plots as well as the EEG chart really allowed me to see the differences between the Alzheimer’s brain and the healthy human brain. This project has taught me a lot about machine learning, statistics, and the brain rhythm biomarkers of Alzheimer’s disease.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=885a1f1464c8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Brainwaves to Insight: Visualizing Rest vs Active Brain States with Electroencephalogram (EEG)…]]></title>
            <link>https://medium.com/@xxxxu0714/from-brainwaves-to-insight-visualizing-rest-vs-active-brain-states-with-electroencephalogram-eeg-20cde812c44b?source=rss-b35b8c371f5a------2</link>
            <guid isPermaLink="false">https://medium.com/p/20cde812c44b</guid>
            <category><![CDATA[brain]]></category>
            <category><![CDATA[brain-computer-interface]]></category>
            <dc:creator><![CDATA[Wennie Xu]]></dc:creator>
            <pubDate>Sun, 02 Nov 2025 03:01:32 GMT</pubDate>
            <atom:updated>2025-11-02T03:06:31.882Z</atom:updated>
            <content:encoded><![CDATA[<h3>From Brainwaves to Insight: Visualizing Rest vs Active Brain States with Electroencephalogram (EEG) data.</h3><p>What if there exists a seamless connection between the human brain and a computer? The thoughts and signals from the brain could be read, interpreted, and transmitted into signals that would be analyzed by a computer. Scary? Well yes of course. But this could hold large potential in the early discovery of brain diseases, allowing for potential treatments to be implemented a lot earlier before the disease progresses.</p><p>This was the logic behind the idea that came out of the blue: What if we can use brain computer interfaces as an early detection sign of Alzheimer’s disease, before any major damages are done towards the brain. More often than not, small changes in the neural networks of the brain come way before any apparent symptoms of memory loss. If we are able to clearly monitor these changes, certain solutions could be implemented before long-term destruction of the brain. This could result in more effective treatment. Read the article here for more details: <a href="https://medium.com/@xxxxu0714/how-can-brain-computer-interfaces-be-used-for-restoring-cognitive-function-in-the-management-and-1792142cd433">https://medium.com/@xxxxu0714/how-can-brain-computer-interfaces-be-used-for-restoring-cognitive-function-in-the-management-and-1792142cd433</a></p><p>However, before any major innovations could be made, a replicate process has to be done in order to understand how brain computer interfaces truly work. In order to do so, I analyzed existing Electroencephalogram (EEG) signals, to differentiate between the resting brain and the brain at task.</p><h3>The logic behind this task</h3><p>In short, Brain Computer interfaces (BCIs) are a sort of technology that harnesses the tiny electrical impulses that occur between neurons during the execution of a task or other cognitive processes. In order to do so, sensors are put either on or inside of the brain to capture this electrical data. The system used in this process is called an EEG.</p><p>In short, this system uses electrodes, tiny metal disks that are placed on the head, to detect voltage fluctuations, recording the difference in electrical potential between that spot on your scalp and a reference point, measuring the relative voltage across regions. In these experiments, markers, which are time-stamp like labels, are implemented in order to indicate what was happening at a specific time in the data.</p><h3><strong>The process:</strong></h3><p>Before any code can be implemented, the data must be found and added. The data used for this project came from the PhysioNet Motor Imagery dataset.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kaCnKYZHOlPYjMayIzy0Wg.png" /></figure><p>A few files from this were then downloaded into Google Co-lab, where the data in the files was graphed out in the picture below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WUZIh6nfRMWuafi7OEXrkg.png" /><figcaption>The raw reading of the EEG headsets, visualized into brain waves</figcaption></figure><p>From this photo, we can observe the visualization of these different brainwaves collected from the different electrodes.</p><p>In this graph, different of the electrodes are shown on the left hand column. The first letter corresponds to what general part of the brain the electrode was placed on (F -Frontal lobe , C- Central sulcus, P- Parietal, O-Occipital, T- Temporal). Depending on the placement on the brain, there could be either 1–2 first letters, with labels with 2 letters meaning that the placement of the electrode are somewhat in between the two respective places. The second letter refers to where on the head, from left to right the electrode is placed. Odd numbers (1, 3, 5, 7) would left hemisphere, even numbers (2, 4, 6, 8) would mean right hemisphere and z refers to the midline (centre of the head). Larger numbers would usually mean that the electrode is placed further away from the centre of head.</p><p>This part of the code already translated the complicated computer signals into a readable graph. Following this, in order to transfer the ideas into actual signals, a few more lines of code are needed.</p><p>The main function of these lines are code are as followed:</p><ol><li>Filter out certain signals that might infiltrate with the brain waves: This helps provide more accurate data. For this specific code, frequencies of 50 Hz and 40 Hz (AC power frequencies) are filtered out in order to clean up interference. The 1–45 Hz power bands are used in order to keep everything within the delta to gamma waves.</li><li>Extracting event markers: There are certain markers in the signal that label specific events. This allows for the places where an action is performed to be identified.</li><li>Cutting the events into epochs: Each epoch would start directly on the event and last 2 seconds.</li></ol><p>The next step would involve extracting and plotting the specific epochs that are marked by “event” or “rest” markers. The result would look something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PzBEHPAOs9-RuvZo-iqvxg.png" /></figure><p>This is a graphical indication of the voltage (μV) collected from each one of these channels, a visualization of the average brainwave activity during that 2 second window. Each individual one of these black lines would indicate data collected from a single EEG channel, with 64 total. They would be placed on different parts of the brain, representing that respective brain activity.</p><p>The final step involves graphing out the data, separating the compiled form into types of brain waves. Processing the data in such way allows for a simple, graphical representation in order to truly understand key differences in the brain during a task or at rest.</p><p>Some of the most important brain waves include the following:</p><ol><li>Theta (4–8 Hz): drowsiness, memory encoding. This is the one most commonly present during sleep</li><li>Alpha (8–12 Hz): relaxed wakefullness, the idle brain. Think of this one as a brain that is awake, but not actively doing something</li><li>Beta (13 Hz — 30 Hz): active thinking, focused</li><li>Gamma (30–45 Hz): sensory processing, high cognition</li></ol><p>The result is as follows:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wKWDEKvNZ6olZMMIos-2Mw.png" /></figure><h3>Major takeaways from this process:</h3><p>This would only be the starting point for something bigger. Understanding the process of brain computer interfaces would be key foundation to potentially using them in solutions for neurodegenerative diseases. Overall, here are my major takeaways from this process:</p><ol><li>In EEG experiments, markers are used as timestamps are used to indicate what was happening in the data</li><li>From these markers, you can pull out specific epochs (periods of time that are usually around 2 seconds) of data that have certain brain waves</li><li>Certain brainwaves would increase when doing certain activities. For example, theta waves increase during sleep, and beta waves increase during focus and thinking.</li><li>Filtering out certain signals of brain waves is vital, as interference may occur.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=20cde812c44b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How can brain computer interfaces be used for restoring cognitive function in the management and…]]></title>
            <link>https://medium.com/@xxxxu0714/how-can-brain-computer-interfaces-be-used-for-restoring-cognitive-function-in-the-management-and-1792142cd433?source=rss-b35b8c371f5a------2</link>
            <guid isPermaLink="false">https://medium.com/p/1792142cd433</guid>
            <category><![CDATA[neuroscience]]></category>
            <category><![CDATA[brain-computer-interface]]></category>
            <dc:creator><![CDATA[Wennie Xu]]></dc:creator>
            <pubDate>Sat, 11 Oct 2025 03:18:03 GMT</pubDate>
            <atom:updated>2025-10-11T03:18:03.205Z</atom:updated>
            <content:encoded><![CDATA[<h3>How can brain computer interfaces be used for restoring cognitive function in the management and mitigation of Alzheimer’s disease?</h3><p>It began subtly: a misplaced key, then a forgotten appointment. But soon, Margaret stares at a photo of her children, knowing she should recognize the smiling faces, yet unable to recall a single one. Birthdays, stories, names, all slip away like water through her hands. With each passing day, another piece of her memory is stolen, leaving her family to watch the person they love slowly fading into a stranger.</p><h3>The memory thief: Alzheimer’s disease</h3><p>60–70% of all dementia is due to Alzheimer’s disease. This is a biological process that begins with the appearance of a buildup of proteins in the form of amyloid plaques and neurofibrillary tangles, causing brain cells to die over time and the brain to shrink. As a result, memory, reasoning, and language skills could be weakened or eliminated completely, behaviour and personality could be changed, and spatial understanding could get altered. Margaret is part of the estimated 57 million people with dementia, with a majority of them due to Alzheimer’s disease.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*A0gktG8uo6PruFK1e9tBsg.png" /></figure><h3>Current treatments for Alzheimer’s disease only target symptoms or delay decline</h3><p>There are currently some sorts of methods to manage the symptoms and target the clumps of the protein beta-amyloid in the brain, the root cause of Alzheimer’s disease. Such a treatment mimics the natural immune response, in order to clear the clumps of beta-amyloid. For example, Lecaneman and Donanemab all are drugs that target mild Alzheimer’s disease and come in an IV infusion. They both come with complications such as infusion reactions, and even brain swelling and brain bleeds. Another solution includes targeting the Tau Tangles in the brain, twisted tau proteins that result in a collapse in brain cell transport. This method is still currently in clinical trials, but it most likely would still also cause a temporary relief of symptoms. Overall, these current solutions are too risky, show only a modest decline, have mixed or failed experimental results, and are not as restorative.</p><h3>The question now arises: How can we truly restore cognitive function for those with Alzheimer’s disease?</h3><p>This is where BCIs (Brain Computer Interfaces) come in.</p><p>Imagine being able to control objects with just your thoughts, sort of like the telekinesis you see in science fiction. In short, this is what a brain computer interface is. Brain computer interfaces, or BCIs, are systems that enable direct communication between the brain and external devices, which allows individuals to control technologies with their thoughts alone. It works by capturing electrical activity generated by neurons in the brain, usually through devices equipped with electrodes. When a person thinks or intends to perform an action, specific patterns of neural activity are produced. The sensors on the Brain Computer Interface could detect those patterns and send a raw signal to the computer for further processing. Then, using Machine Learning, the Computer is able to associate certain brain wave patterns with an action, allowing it to execute certain commands such as being able to move a cursor or a robotic arm.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/632/1*BjUKRALPE4NaNbqZsqtigA.jpeg" /></figure><h3>The two main types of Brain Computer interfaces are invasive and non-invasive. As suggested by the name, the difference between the two comes in the implementation of the device on the patient.</h3><p>Invasive BCIs involve the electrodes being implanted directly into the brain tissue, which yields high resolution signals but also require surgery and have medical risks. Electrodes are surgically implanted either within the gray matter (microelectrode arrays), on the cortical surface (ECoG grids) or into deeper brain structures (deep brain stimulation, DBS, electrodes). The information collected could be as precise as being able to detect the electrical activity of individual neurons or small groups of neurons, yielding both high spatial and temporary resolution signals compared to non-invasive methods.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zbh_YfuGlo4SnF97UdHwtQ.png" /></figure><p>Non-invasive BCIs, on the other hand, are placed on the scalp, allowing it to detect brain activity without a surgery. Though the signals offered are less detailed than those of an invasive BCI, it is a lot safer and suitable for general use. These usually come in the form of an electroencephalogram (EEG), using electrodes placed on the scalp in order to record brainwave patterns.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ttG0QACh0yHJ9nn0IzXZ4Q.png" /></figure><h3>How do these sci-fi-like machines help with Alzheimer’s disease?</h3><p>BCIs have promising potential in managing Alzheimer’s, but they are still largely experimental and not widely used as a standard treatment.</p><h3>Cognitive Enhancement is possible with BCIs</h3><p>The use of a BCI could possibly have the ability to change the makeup of the brain, resulting in a boost in healthy brain function. For example, one of the key ways that this could happen is by enhancing memory. BCIs may have the ability to strengthen memory encoding by delivering precisely timed stimulation to the hippocampus or prefrontal cortex, parts of the brain that are involved in the formation of memory. In early human trials, this improved short term memory performance by up to 35% in some of the patients. This is alike to giving the brain a “save” button that reinforces neural patterns.</p><p>BCIs could also modulate oscillations like theta and gamma waves, helping to optimize how different brain regions communicate. A healthy brain has neurons communicating to each other using rhythmic patterns of electrical activity, such as theta waves (linked to memory encoding), and gamma waves (linked to learning and integration). In Alzheimer’s, these rhythms become disrupted, causing a decrease in how the brain forms memories and learns. BCIs could be used to train these rhythms with very precise stimulation, allowing the efficiency of information transfer between different regions of the brain. This could potentially result in faster memory consolidation and smoother communication across neural networks.</p><p>Furthermore, learning depends on synaptic plasticity, the ability of connections between the neurons that can strengthen or weaken over time. BCIS could hence be used to deliver stimulation at the exact moment new information is being processed, which helps reinforce those connections. A process known as Spike Timing Dependent Plasticity (STDP), controlling and adjusting the precise timing of the electrical impulses of the presynaptic neuron to be before the postsynaptic neuron, the connection between them could strengthen, resulting in higher synaptic plasticity, therefore a strong brain over time.</p><p>In short, in the direction of Alzheimer’s disease, BCIs could act like a neural accelerator, aligning rhythm and boosting plasticity, which could result in the restoration of memory rhythms, enhancing encoding and recall, and compensating for damaged circuits.</p><h3>Monitoring Alzheimer’s progression with BCIs as an important biomarker of the development of disease.</h3><p>Another angle where BCIs could be implemented is in the monitoring of the progress of Alzheimer’s disease in real time. This is an important aspect, as cognitive tests or MRI scans could only observe the development every few months.</p><p>Alzheimer’s disease disrupts the communication between the hippocampus, prefrontal cortex, and parietal lobes, all places that are vital for encoding and recalling information. More specifically, the theta and gamma wave activity gets disrupted. Hence, monitoring the strength and the synchronization of such rhythms could be important in detecting the progression of this disease. As changes in oscillations appear before symptoms worsen, this could be used as an early biomarker in order to help manage the disease from progressing at an earlier stage.</p><h3>Existing evidence and experimentation shows that BCIs have the potential to manage the progression of Alzheimer’s disease.</h3><p>BCI-based cognitive training has been proven to significantly enhance cognitive functions in dementia patients over the span of six weeks, outperforming traditional interventions like music and physical exercises. As well, there has been a reported that BCIs have a high accuracy (86.47%) in distinguishing early Alzheimer’s Disease from mild cognitive impairment and healthy states, aiding in early detection.</p><h3>The DARPA RAM program shows promising results in restoring normal memory function to those suffering from brain injury or illness using BCIs</h3><p>An existing approach to mitigate the effects of memory loss is the Restoring Active Memory (RAM) program, aiming to mitigate the effects of brain injury in military Service members. The condition results in an impaired ability to retrieve memories formed prior to injury, and a reduced capacity to form or retain new memories following injury, much like Alzheimer’s. This entire neural interface is based on building multi scale computational models with high spatial and temporal resolution that describe how neurons code declarative memories. In other words, it is studying how neurons are able to restore memories that can be consciously recalled and described in words. New methods for analyzing and decoding neural signals are also used to understand how targeted stimulation might be applied to help restore function to the injured brain.</p><p>This foundational work led them to develop implanted closed loop systems that can deliver targeted neural stimulation to restore normal memory function. This is placed with electrodes on the hippocampus, the part of the brain that encodes memory, in order to selectively stimulate parts of it allowing it to help overcome the damage caused by trauma.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uLgon7F-tl7qczUUEBMKbw.png" /></figure><h3>Despite promising results, certain barriers do exist before the implementation of such a solution.</h3><p>Many barriers and technical limitations exist with BCIs that might pose a significant barrier before this technology could be implemented in the treatment of Alzheimer’s.</p><h3>Signal Quality and Noise</h3><p>The entire technology of a BCI comes with picking up signals and transmitting them to a machine. However, non-invasive BCIs often produce weak and noisy signals because electrical activity has to pass through the skull and the scalp. This makes it a lot harder to measure deeper structures in the brain, such as the hippocampus, which is absolutely critical in Alzheimer’s disease. On the other hand, though invasive BCIs give clearer signals, they require risky brain surgery, which might not be safe for elderly patients.</p><h3>Long Term Reliability</h3><p>Implants also have the risk of scarring brain tissue or degrading over time, reducing effectiveness and potentially causing damage to the sensitive brain structures. Electrodes could have the possibility of corroding, breaking, or losing conductivity. As well, the brain treats implants as a foreign object, which causes glial scarring around the electrodes. With scar tissue insulating the electrodes, signal detection and stimulation are dampened drastically.</p><h3>Variability</h3><p>As well, since the patient’s brain is constantly changing and evolving over time, constant calibration of the equipment, which could be technically challenging and quite costly for the patient.</p><p>There is also an issue of adaptation. Alzheimer’s disease looks different for everyone: different people have different amounts of damage in different parts of the brain. In order to apply this technology to everyone, each BCI must be specially adapted and calibrated to fit the individual needs.</p><h3>Accessibility</h3><p>Cost and the constant need for recalibration limits these devices to only people who are quite wealthy, as the fees for a brain computer interface with the potential of going deeper into neural networks are set to be quite expensive. This limits the accessibility of these devices, leaving a larger population of people with Alzheimer’s out of the treatment question.</p><h3>Future Directions</h3><p>In conclusion, Brain computer interfaces hold exceptional promise in transforming how we detect, understand, and treat Alzheimer’s disease. By directly integrating in neural circuits, BCIs could not only slow disease progression through early detection but could also restore aspects of memories and cognition that were once thought to be permanently lost. However, the path forward is complex: as there are many technical hurdles such as long term reliability and safety. Much research is needed to understand the human brain in order to minimize invasiveness, improve signal accuracy and reliability, and achieve long term clinical safety.</p><p>The rapid pace of innovation suggests that BCIs could one day shift the current state of Alzheimer’s care from merely just managing decline to actively rebuilding cognitive function. In doing so, it offers patients not just more years of life, but more years of life truly lived.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1792142cd433" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>