Reinforcement Learning Creates Human-Like AI

Dr. Norberto Grzywacz, a very talented and impressive professional from Loyola University Chicago, gave a talk at the Neuroscience Seminar about his research related to reinforcement learning on social polarization. Our values is how we make decisions and these values come from the society we live in. As life goes on we are constantly receiving feedback from others. When we do something others like, we get some kind of positive feedback. While the reverse is also true, when we do or say something others do not agree with we receive negative feedback. This shapes our future decisions and actions. This is called reinforcement learning. Reinforcement learning can lead to polarization because we like the feeling of positive feedback. Therefore, we interact with the people that give us this feeling, and the loop continues until we are only interacting with people that are the same as us, creating a false reality that everyone else is exactly like us or should be. Something striking that Dr. Grzywacz mentioned was that artificial intelligence (AI) learns in this same manner.

This article written by Julian Horsey delves into more insight of how AI learns in the same way and how human input impacts that learning. An already trained AI is used in this process of reinforcement learning from human feedback (RLHF). That means one that already has a large amount of data and understands the foundations of human language. First it is fined tuned to desired behaviors by giving the AI examples to guide it to correct responses. Secondly, humans help the AI learn. This is done as AI completes tasks, someone assigns the response with a corresponding number to label how desired the response was. This is reinforcement learning in the same way humans learn as a higher number equals a better response. Finally, the AI maximizes the desired responses based on the numerical values assigned earlier. However, RLHF has some drawbacks. The main ones are overfitting and bias. These lead to polarization within the machines themselves. As humans who do the numerical values are inherently biased, the machines can begin to give too narrow of responses or perpetuate stereotypes and malicious biases inadvertently. RLHF does have its benefits of generously increasing AI’s ability and scope to perform tasks in a more naturalistic manner with more nuanced communication. However, RLHF must be done carefully, just like in humans RLHF can lead to polarization and harmful outputs, but if done right it can lead to improved AI that can maximize its benefits to society as a whole.

Reinforcement learning is an incredibly interesting topic of how the brain works to belong within society and be “one of the pack”. However, it also highlights how humans can easily fall into the extremes. It is fascinating how AI can emulate this behavior already acting more human-like than ever before, signifying that we must remain cognizant of what AI is becoming and the implications that it can provide for our society both beneficial and detrimental.

 

References:

Grzywacz, N. M. (2025). Comparison of distance and reinforcement-learning rules in social-influence models. Neurocomputing, 649, 130870. https://doi.org/10.1016/j.neucom.2025.130870

Horsey, J. (2024, August 8). AI reinforcement learning from human feedback (RLHF) explained. Geeky Gadgets. https://www.geeky-gadgets.com/reinforcement-learning-from-human-feedback/

When Dreams Talk Back: Rethinking Cognition Across Sleep and Wake

            Dr. Gabriella Torres-Plata's research this semester challenged long-held sleep science assumptions and presented evidence that contradicts them. The most significant assumption she challenged was that dreaming could only be studied after the fact. In place of that, Torres-Plata and colleagues have shown that some people who become aware while they are dreaming (lucid dreamers) can receive information and respond while still in REM sleep - they can answer yes/no questions, do basic arithmetic, and identify different sensory inputs based on coded eye movements and slight facial expressions. Torres-Plata and colleagues have opened the door for studying what happens in the mind while we dream, including memory creation, emotional processing and creative problem solving. The previous method of studying dreaming has been limited to the degree to which a person can remember what happened while they were asleep. This study has the potential to help us understand how the brain works at all times, and to make better use of our ability to dream. 

The concept that Torres-Plata explored has additional support in a recent paper published in the journal Neuron titled "Linking Memory and Imagination Across Sleep and Wake States" (Wamsley et al.), that came out in 2024. While Wamsley and his colleagues did not explore dream dialogue, they did find similar brain activity in REM states (where dreams occur), and in states of daydreaming. They were able to observe the way that the brain changes old experiences into something new, using areas of the brain such as the hippocampus and cortical networks. Both processes involve flexible, creative recombination of past experiences, guided by the hippocampus and association cortices. The Wamsley et al. study suggests that dreaming is not just “random noise,” but part of a broader system for simulating, reorganizing, and sometimes inventing mental content. 

Ultimately, the possibility of having a real-time conversation with someone while they are dreaming may provide the opportunity to study how the process of transforming memories occurs on a moment-by-moment basis. However, it also raises many important issues related to privacy, autonomy and the extent to which researchers should be allowed to intrude upon an individuals internal mental processes. Real-time dialogue might eventually allow researchers to probe how memory transformations unfold moment by moment. At the same time, these possibilities bring up important concerns about privacy, autonomy, and how much access researchers should have to an individual’s inner cognitive world. 

Both the seminar and the Neuron article highlight REM sleep as an active and dynamic state—not a pause in consciousness, but a unique mode of thinking. Together they suggest that dreams may be a window into how we imagine, problem-solve, and reshape our emotional experiences, both during sleep and while awake. 

References 

Konkoly, K. R., Appel, K., Chabani, E., Mangiaruga, A., Gott, J., Mallett, R., Caughran, B., Witkowski, S., Whitmore, N. W., Mazurek, C. Y., Berent, J. B., Weber, F. D., Türker, B., Leu-Semenescu, S., Maranci, J.-B., Pipa, G., Arnulf, I., Oudiette, D., Dresler, M., & Paller, K. A. (2021). Real-time dialogue between experimenters and dreamers during REM sleep. Current Biology, 31(7), 1417–1427.e6. https://doi.org/10.1016/j.cub.2021.01.026 

Wamsley, E. J., Elfassy, N. M., Ambo, L. K., Wangsness, H., Mughieda, N., Jensen, O., & Bhattacharya, M. (2024). Linking memory and imagination across sleep and wake states. Neuron, 112(9), 1501–1518. https://doi.org/10.1016/j.neuron.2024.02.012 

Dreaming Circuits and Directed Memories: How Selective REM Control Sheds Light on Sleep’s Hidden Functions

    From this semester, the one striking idea portrayed was the discovery that the lateral posterior nucleus (LPN), a higher-order thalamic visual nucleus that takes part of a specific role in REM sleep regulation. Guerrero and Cavanaugh’s conduct of this study demonstrates that silencing the LPN selectively leads to the reduction of REM sleep in mice. In a way, initiating this method leaves non-rapid eye movement (NREM) sleep almost intact. Going along the same idea, selective circuits lead to governing sleep function, which is applied to NREM sleep, the critical phase of deep memory consolidation. This proposes the fact that REM may rely on targeted support from sensory-related thalamic circuits instead of whole-brain sleep mechanisms. In another sense, REM isn’t just “dreaming turned on” but appears to be an adapted situation, in which the neural structures are working all around and stepping in at the right time needed. This finding raises the question of whether REM is supported by specialized neural circuits, and could altering REM change memory, emotion, or even dream content?

    A 2021 study from Northwestern University was widely covered in news outlets such as NPR and Science Daily. They examined the occurrence called target memory reactivation (TMR), which is the process of activating the brain with keen sounds during sleep to either strengthen dreams or to modify them. With this study presented, the participants practiced a melody on a keyboard while also listening to a specific sound. The scientists replayed that cue during sleep, which led to the factor the participants performing the melody significantly better. The results acquired from these participants the next morning were better than allowing that cue to be replayed overnight. The key point of this research displayed how specifically REM and light NREM stage sleep are not submissive, therefore is considered an active state. That is because the specific brain circuits present selectively alter what we perceive and how well we remember it. This study resonates with Guerrero and Cavanaugh’s work. If REM sleep needs the LPN to maintain its normal structure,  then damaging that specific circuit could interfere with the neural cycle that TMR relies on. There could be an indication of an “entry point” for comprehending how sensory information, dreams, and memory consolidation could interact at a circuit level. 

    In the TMR study done at Northwestern, memories are reinforced by activating certain cortical pathways during sleep. Looking into the mouse in the REM study, by selectively silencing a visual thalamic structure, it weakened REM, which supports memory consolidation. There is the idea of distinct brain circuits that may play a role, determining what kinds of information get processed during sleep. These studies have opened the door to a new interpretation of sleep, how it's a series of regulated brain modes in which each is linked and supported by distinct neural structures with special functions. Sleep ultimately becomes less like a blanket covering for the brain but more like a series of subtly tuned spotlight beams that illuminate the different neural pathways, all at different times. As the continuation of learning proceeds, on how sleep interacts with memory and perception. The idea that one day we may influence how we dream seems to be realistic, not just being able to know whether we dream or not. The conclusion from the studies of Guerrero and Cavanagh suggests how these possibilities begin in the cortex but also how surprisingly powerful thalamic circuits can shape the landscape of our sleeping minds.

References

Guerrero, J.R., & Cavanaugh, J. (2025). Silencing the lateral posterior nucleus produced a highly selective reduction in mouse REM sleep. Neurobiology of Sleep and Circadian Rhythms. No DOI available. https://www.sciencedirect.com/science/article/pii/S2451994425000136  

Anthony, J. W., Cheng, L. Y., Brooks, P. P., Paller, K. A., & Norman, K. A. (2021). Sleep spindle refractoriness segregates periods of memory reactivation. eLife, 10, e70068. https://doi.org/10.7554/eLife.70068 

How People (& Values) Learn Like Machines by Gianna Eisen

Have you ever noticed yourself distancing yourself away from a person who has widely different opinions from you, or get tired of explaining yourself for your opinion? What about finding someone with whom you get along and who shares similar opinions to you? While catching up with friends or scrolling through comments on a social media post, it is easy to notice how clusters of opinions are formed and how individuals create a social group for themselves based on shared beliefs. But how does our brain perceive this polarization? Norberto Grzywacz, in his study Comparison of Distance and Reinforcement-Learning Rules in Social-Influence Models, studied how decision-making and social influence could follow the same learning rules that direct artificial intelligence. In past research, models of social influence automatically assume that individuals are swayed by those who already think like them, which is where the ‘distance-based’ rule, where influence takes precedence as opinions diverge. However, in Grzywacz’s research, we think about whether our brains use reinforcement learning, which is the same process used in AI systems. This allows people to learn which values or opinions ‘ pay off’ socially and show that behaviors that are rewarded (like by acceptance or appraisal) are then strengthened. Grzywacz’s simulation showed that both systems create polarization, but reinforcement learning produces realistic group patterns, like groups aligning or dividing suddenly due to unforeseeable shifts.

This idea is closely connected to a 2020 study by Levorsen et al., Testing the Reinforcement Learning Hypothesis of Social Influence. By using fMRIs and computational modeling, researchers were able to assess if social conformity, the tendency to match others’ opinions, operates on the same neural mechanism as reinforcement learning. Participants received both social feedback, on how their opinions compared to others, and reward feedback, monetary gains or losses. According to the study, the brain didn’t treat these two forms of feedback the same way, showing that while reinforcement depends on reward-processing regions of the brain, like the striatum, social influence, on the other hand, engaged additional areas related to social reasoning and self-evaluation. These findings show that Grzywacz’s reinforcement-learning model captures many aspects of social polarization, but real-world influence could be even more complex. Our brains don't just chase rewards, but respond to social norms or identity. This idea confirms that polarization does not only occur because opinions are ‘rewarded’ but because it defines who a person is within a group. Both Grzywacz and Levorsen’s findings show how individuals learn from one another. Rules like reinforcement learning can help to explain why opinions spread and cluster, but the emotional and social context of the opinions themselves makes the process human. Understanding both sides, the algorithmic and the social, can help our machines, like AI, learn values using reinforcement like systems, that pick up our biases and polarization patterns. Designing systems that can distinguish between social learning and reward learning can help prevent the divisive act of polarization.

Works Cited:

Grzywacz, Norberto M. Comparison of Distance and Reinforcement-Learning Rules in Social-Influence Models, Neurocomputing, Volume 649, 2025. https://doi.org/10.1016/j.ne ucom.2025.130870

Levorsen, Marie et al. “Testing the reinforcement learning hypothesis of social conformity.” Human brain mapping vol. 42,5 (2021): 1328-1342. doi:10.1002/hbm.25296

Crossing the Line: How our Brain Keeps Vision Seamless and Why Symmetry is Key.

 Most of us have never thought about the center line of our vision or actively seek it out. You look to your left, you look to your right, and even so the world manages to keep is balance and stay in one piece. In a recent article by MIT News, “How the brain splits up vision without you even noticing”, tells us why. As something moves across your view, the brain quietly passes the information from one hemisphere to the other so there isn’t a noticeable glitch or gap. For example, if runners are running a relay race and one passes the baton to the next runner up, and they match their speed so the baton is handed off cleanly and the race will continue without a hitch. This speaks to how perception works in everyday life, such as walking down a crowded sidewalk, during a sporting event, or even while doom scrolling through reels. 

In our Neuroscience seminar, our reading “The role of vertical mirror symmetry in visual shape detection” by Machilsen et al. (2009) talk about how the left and right side of our vision matching isn’t a design for aesthetic purposes but rather helps our brain find shapes in clutter. In the experiment from the reading, participants were asked to spot outline of shapes made from tiny line fragments with slightly turned little lines that made the outline so the shape was harder to see. Through the multiple conditions, shapes that were symmetric were much easier for the participants to decipher compares to chapes that were symmetric. This shows that symmetry acts as shortcut for when scenes aren’t clear or messy for our brains to outright decipher, so our brains use the other half that is clear or more recognizable to fill in the other half. 

But how do these ideas work together? The article by MIT explains how features slide across our vertical midline and the two hemispheres of our brain work together to keep it a continuous scene. Since vertical symmetry is connected by the midline, our brains are able compute that this half should mirror the other half. Symmetry pops because our brain is designed to create one big picture even with many moving pieces. This also confirms what Dr. Baker, our guest speaker, mentioned in his talk. Our vision is built up of many little pieces that work together to create the big picture. 

References: 
https://www.luc.edu/psychology/people/facultyandstaffdirectory/profiles/bakernicholas.shtml

David Orenstein  |  The Picower Institute for Learning and Memory. (n.d.). How the brain splits up vision without you even noticing. MIT News | Massachusetts Institute of Technology. https://news.mit.edu/2025/how-brain-splits-vision-without-you-even-noticing-0926

Machilsen, B., Pauwels, M., & Wagemans, J. (2009). The role of vertical mirror symmetry in visual shape detection. Journal of Vision, 9(12):11, 1–11, http://journalofvision.org/9/12/11/, doi:10.1167/9.12.11. 

Ultraprocessed Foods Linked to Early Symptoms of Parkinson’s

 In class, we had the opportunity of having Dr. Mary Makarious as one of our guest speakers to talk about a program where she takes participation in to have an explanation towards Parkinson’s. Some early symptoms to detect in suspicion of early Parkinson’s are shaking unexpectedly, unable to sleep, losing sense of smell, not being able to stand properly, slow movement, expressionless face, etc. Parkinson’s disease is a disorder centralized of the nervous system that affects movement which over time unfortunately only gets worse. In a more scientific explanation, the dopamine in your brain stops functioning as the cells stop making it.  Global Parkinson’s Genetics Program (GP2), is made from different researchers around the world that do different research to find a cause for Parkinson’s in a genetic aspect. They give us a broader look into how different populations from around the world come to face this disease differently. 

I have come across an article, “Ultraprocessed Foods Linked to Early Symptoms of Parkinsons’s” which center a recent study on the potential cause on how our food choices can have an impact on brain health. With studies, you are going to have successful and Non successful ones, and this one came out to not prove that foods can be cause of this disease but did help on giving us more insight towards food related to Parkinson’s. This involves refraining from consuming too much processed foods, as those who consume it most have been revealed to be at higher risk of developing symptoms than those who consume less. Not only does it affect Parkinson’s, but it can have other life-threatening health problems like cancer, diabetes, dementia, heart disease, etc. This research developed in China and the US and involved analyzing the diets of their participants over years to find that ultra processed food takes a big toll on neurological diseases.  

Although there is still a long way ahead to find a true cause and even a potential solution to this life-threatening disease, research like the one from Dr. Makarious and the one previously mentioned in the article, all contribute to an effort for answers. Many people living with this disease daily are fighting for their lives, and these researchers are putting in an effort to show that hope has not been lost. 

Reference:  

Callahan, Alic. Ultraprocessed Foods Linked to Early Symptoms of Parkinson’s. The New York Times (2025). https://www.nytimes.com/2025/05/07/well/eat/ultraprocessed-foods-linked-to-early-symptoms-of-parkinsons.html?smid=url-share 

Understanding the World of Silence

     In the beginning of the semester, Dr. Wei-Ming Yu presented the concepts of Hepsin- which is an important transmembrane serine protease, TMPRSS1, that is crucial for the proper functioning of the cochlea and tectorial membrane. To be better prepared for this presentation, we had to read the article "Critical role of hepsin/TMPRSS1 in hearing and tectorial membrane morphogenesis: Insights from transgenic mouse models" by Yang, Ting-Hua et al. This article first started by introducing that mutations in a type II membrane serine protease in family members are highly associated with a non-syndrome hearing loss although some of those mechanisms are still unclear to be understood.  

First and most importantly, Dr. Yu explained the different mechanisms of how sound is produced by vibrations that create particles in a surrounding medium such as air.  This in turn result in a wave of vibrations that travel through air to the eardrum. When sound enters the ear, a series of hair cells activate in which fluid waves bend hair cells called cilia inside the cochlea. Vibrations travel through the ear canal then press against the eardrum, causing it to oscillate, or move back and forth in a rhythm., The movement of fluid bends the cilia in the cochlea which converts mechanical energy into electrical signals that are processed as sound perception via the brain.

Dr. Yu explained in depth at the molecular level that the proper technique of the tectorial membrane is important for accurate the stimulation of these cilia. This reveals that hepsin/TMPRSS1 plays a key in how we hear sounds. In transgenic mouse models lacking hepsin, the tectorial membrane was abnormally shaped and detached from the hair cell leading to severe hearing impairment. This proved that hepsin is necessary for maintaining the correct composition and structure of the tectorial membrane through processing of some specific extracellular matrix proteins.

In the article “Noise Exposures Causing Hearing Loss Generate Proteotoxic Stress and Activate the Proteostasis Network” by Ramirez et al. describes how exposure to loud sounds damage to the inner ear and imbalances the cochlea. They experienced on rats by exposing them to different levels of sound: moderate, loud and very loud studying the biological effects of noise induced hearing loss. The results were that they found loud levels of sound causes protein in the cochlea to unfold and become damaged. 

Even though both articles are different from one another, they explain concepts at the molecular level of hearing loss providing evidence of how environmental and genetic factors produce sound. 

 

1.     Noise Exposures Causing Hearing Loss Generate Proteotoxic Stress and Activate the Proteostasis Network

Jongkamonwiwat, Nopporn et al. Cell Reports, Volume 33, Issue 8, 108431

2.     Yang, T. H., Hsu, Y. C., Yeh, P., Hung, C. J., Tsai, Y. F., Fang, M. C., Yen, A. C. C., Chen, L. F., Pan, J. Y., Wu, C. C., Liu, T. C., Chung, F. L., Yu, W. M., & Lin, S. W. (2024). Critical role of hepsin/TMPRSS1 in hearing and tectorial membrane morphogenesis: Insights from transgenic mouse models. Hearing research, 453, 109134. https://doi.org/10.1016/j.heares.2024.109134