Seeker - Blog , Neuroscience

Depression, habenula, and the pitfalls of setting rigid goals

Sun, 30 Aug 2026 22:17:59 +0530

I recently visited my cousin after eons and had the fortune of spending some quality time. My cousin, who is about 17 years old, is full of life and quite chatty. Naturally, we discussed all sorts of things. During the conversation, he excitedly began sharing details about his fitness journey. He told me that he had started hitting the gym a few weeks ago and took great pride in the fact. While we were engaged in conversation, his mom interrupted us and began expressing her concerns over his behaviour.

My aunt told me that my cousin’s behaviour of late had been annoying her. When I asked what frustrated her, she remarked on how my cousin had refused to visit his grandmother on her birthday because he wouldn’t be able to meet his protein requirements, and how he had got riled up when she insisted that he join them.

Now, he interrupted his mom and started bragging about how consistent he had been. He told me that he did not want to skip his daily routine - which included going to the gym, walking a certain number of steps, eating a certain amount of protein, and so on - and that missing any of these things made him angry.

While I was genuinely happy for him - after all, he was trying to set higher standards and build healthier habits - his rigidity unsettled me. It reminded me of a pitfall that, based on my reading of Dr. Kyra Bobinet’s The Unstoppable Brain, he might be at risk of falling into, which is setting SMART goals (counting steps, calories, protein intake, tracking diet etc) that become unsustainable and defining success so narrowly that any deviation feels like failure.

Note: This is not a preachy post, nor is it a recommendation. I’m merely sharing Dr. Bobinet’s views on the pitfalls of SMART approaches to behaviour change, because I echo her sentiment and believe it's useful to know her view.

Before I continue, here's a quick intro of Dr. Bobinet taken from her website.

Kyra Bobinet, MD, MPH, has passionately pursued and studied the truth about behavior change for nearly three decades as a physician, public health leader, healthcare executive, and behavioral expert. An award-winning health innovator and thought leader, Dr. Bobinet has an MD from UCSF School of Medicine and an MPH from Harvard University. She belongs to the Stanford Medical School AIM lab, where she has taught on health behavior change, and is CEO/founder of Fresh Tri, a behavioral software company based on the latest neuroscience of habit formation and lasting change.

Let's admit - building a new habit is difficult, because it requires consistency and

discipline. Setting SMART goals is usually considered a good way to develop these (we'll explore later why). But Dr. Bobinet draws on her decades of experience as a behaviour-change expert to argue otherwise. In Unstoppable Brain, she contends that SMART approaches to habit building are performative in nature and argues that such approaches work only in the short term and can, in fact, be detrimental for long-term behaviour change.

To understand why they are not just ineffective, but often detrimental w.r.t long-term behaviour change, it's important to learn the neuroscience undergirding these approaches.

Dopamine & behaviour change

SMART goals are attractive because they are measurable, concrete, and directional. Pursuing these goals engages our reward system - dopamine pathways.

You set a goal, move towards it, and experience a sense of reward upon achieving the goal. At the end, you feel good about yourself, which makes you "learn" the behaviour. Repeating this enough times is likely to turn the behaviour into a habit.

Much of the discussion around behaviour change and habit formation has therefore focused on rewards, incentives, and the anticipation of achieving a desired outcome as a way to motivate behaviour.

But what's likely to happen when you don't achieve the goal? In other words, what happens when you fail, or rather, "perceive failure?" That's the very question Dr. Bobinet draws our attention to. Let's understand this by looking at a gut-wrenching experiment conducted on dogs by psychologists in the 1960s.

Learned helplessness

In the 1960s, psychologists Martin Seligman and Steven Maier conducted a series of experiments in which dogs were enclosed in a kennel and given electric shocks. Initially, the dogs squealed in pain, frantically moving back and forth and attempting to escape the cage, but after multiple attempts, they eventually stopped trying. Later, the scientists made it possible for the dogs to escape. What surprised them was that despite being in a situation to escape, many of the dogs simply gave up. Their earlier experience seemed to have taught them that there was nothing they could do to change their situation.

The phenomenon came to be known as learned helplessness. In the above example, the dogs learned to be helpless. This occurs when an individual repeatedly faces a negative or uncontrollable situation and stops changing their circumstances, even when they seemingly can do so. Dr. Bobinet calls this the know-do gap, which takes the following form

"I know what I should do, but I don't do it."

"Nothing I do makes a difference." etc

This is one of the signs of depression.

What's driving this attitude? THE HABENULA

Note: She highlights multiple studies like this one that discuss the role of habenula in depression.

Habenula, the motivation kill switch (borrowing her term)

Habenula is an ancient structure located in brain's epithalamus region. According to Dr. Bobinet, habenula is perhaps the most important controller of behaviour, and it's activated whenever we perceive failure. Since these ideas form the crux of my post,

I'd like to quote a few passages from the book in the sections that follow to convey Dr. Bobinet's message with fidelity.

"A half-centimeter area of our brain's epithalamus, this potent neuro-anatomical mechanism activates whenever there's perceived failure and then, often subconsciously downregulates one's motivation to try again"

PG 48
"When activated, habenula is a kill switch for our motivation. We stop trying new experiences or building healthy habits whenever we think we failed at, say, sticking to a diet or even controlling our anxiety."

The below is perhaps the most important insight of all.
PG 48
"Keep in mind that there is rarely absolute or terminal failure; mostly we experience perceived failure in terms of our behaviour change. But to habenula, they are the same thing."

Ex:Assume you set your mind on completing 10k steps everyday of the week. Now, being able to complete only 7k on a particular day is not a failure in any absolute terms. It's a failure only in relation to your goal of completing 10k everyday.

Dopamine & Habenula

Pg 49,

"The habenula's power over your behaviour is best illustrated by imagining that you are sitting in a car. The gas pedal represents your motivation and the concomitant dopamine that rewards you for pursuing goals. In fact, this "gas pedal" has been the focus of health improvement programs.............But what we all missed is - the brakes were on! The habenula acts like your brake pedal. And just like in a real car, you can floor the gas all you want, but if the brakes are on, you're not going anywhere. In other words, an activated habenula has more power over your behaviour than an activated reward system."

Essentially, failures register more strongly and painfully than successes feel rewarding. As a result, repeated failure to achieve a goal can undermine motivation far more powerfully than sporadic success can sustain it.

Note: Incidentally, I am reminded of Daniel Kahneman and Amos Tversky's loss aversion theory. Loss aversion is a psychological bias in which people prefer to avoid losses more than getting equivalent gains.

In her view, setting SMART goals often leads to relapse in behaviour because our environment can't be programmed to perfection every single day. Some days, you are sick. Another day, you have a commitment or a function to attend - as in my cousin's case. On some other day, you have less energy, etc. Therefore, a rigid mind that looks at success and failure as binary (10k steps success, anything less, I failed), is likely to suffer more - this can range from not pursuing the behaviour (lack of motivation) to becoming depressed (the detrimental bit).

Adapting to change, staying flexible, and iterating on our processes are key to long-term behaviour change, according to Dr. Bobinet. She has an entire chapter devoted to discussing these.

Conclusion

We are complex creatures and our brains work in fantastically complex ways. Neuroscience is still an evolving field, with scientists constantly coming up with new theories, experiments, and models to unpack the mysteries of our brain. Whenever new theories with better explanatory power and experimental evidence relating to a phenomenon emerge, it's worth paying attention to them. This in no way implies that the newer theory has the final word w.r.t explaining the phenomenon.

This is not limited to Neuroscience, but rather to the scientific enterprise as a whole. [Atomic theory, standard model, big-bang cosmology, computational theory of mind etc]. You name it.

That's the spirit in which we must construe Dr. Bobinet's take about habit change and behaviour formation as well. And she's not even decrying SMART approaches in a blanket manner.

In her own words. The below quote captures the essence of her argument.

"I personally have nothing against goals, awards, or performances. I celebrate incredible human achievements and breathtaking feats as much as the next person. But performance mindsets, if used inflexibly, for the same individual or organization, have been proven to cause harm eventually. This is because they are best for sprinting but not for marathoning."

If SMART approaches to building habits have worked for you, there's perhaps no reason for you to change. But if they have not worked, there's value in understanding Dr. Bobinet's view and developing an iterative mindset. At the end of the day, if we start enjoying the process, without having rigid definitions of success or failure, we are likely to develop healthier habits and foster long-term behaviour change.

Get Started Now

How experiments in AI influenced our understanding of the human brain

Mon, 29 Jun 2026 13:07:13 +0530

I am currently reading Max Benett's "A Brief History of Intelligence" - which is profoundly engaging - in which Benett charts the major evolutionary breakthroughs in the development of intelligence. With each chapter, he offers fascinating insights on how intelligence, as we understand today, emerged over time, and in the process, draws parallels between human intelligence and that of AI. Consequently, he records the progress made so far with AI and highlights the pitfalls we may have fallen into in pursuit of achieving human-like intelligence.

The first breakthrough was steering - the ability to move towards a target (actively hunt), aided by bilateral symmetry. However, I'd like to discuss "reinforcement learning", the second breakthrough in this post, because it underscores how experiments in AI have also influenced the study and understanding of the brain.

The relationship between AI and neuroscience is often imagined as a one-way street, where understanding the brain leads to the development of AI. But here is an example of how building machines helped scientists better understand the human brain. It’s an interesting story worth sharing.

The second breakthrough : reinforcement learning

After steering, the second breakthrough in our evolutionary history of intelligence was reinforcement learning, a capability developed by the first vertebrates. But first, what is RL?

You can define it in multiple ways, but at a basic level, reinforcement learning refers to an animal's tendency to pursue behaviors that lead to rewards and avoid behaviors that don’t based on trial-and-error learning. Scientists like Marvin Minsky and Edward Thorndike believed that this mechanism was fundamental to how humans and animals learn.

The hypothesis was simple - an animal is likely to repeat a behavior that produces a satisfying outcome, while not repeat a behavior that produces a discomforting outcome. Therefore, the behavior that leads to a reward gets reinforced (the animal learns to repeat it), while the behavior that doesn’t gets omitted.

As scientists developed AI systems, they believed they could rely on the same seemingly straightforward logic - an AI should strengthen behaviors that lead to rewards and weaken those that do not.

Minsky built SNARC, the world's first neural network based on this principle. However, SNARC started failing miserably in completing even simple tasks or winning simple games. This left Minsky and other scientists of the day confused. Trial-and-error learning seemed intuitive and straightforward. Yet it was barely working in machines. However, they soon realized the problem.
The temporal credit assignment problem

The challenge was in determining which behavior deserved credit or must be reinforced?

Note: The author gives different examples in the book. I'll give a simple one.

Consider a dog that must perform a sequence of 7 actions before receiving a reward. It must first press a lever, then run through a tunnel, and perform several other actions before finally choosing the green ball over red to obtain a piece of chocolate.

Assume that the final action is reinforced - (dog picking the green ball over red), because it immediately preceded the reward. The dog would never have reached the final stage if it had not performed the earlier correctly.

For example, if the dog had not pressed the lever at the beginning, it might never have been able to enter the tunnel and complete the remaining steps. In that case, the first action was just as important as the last. If the first action is reinforced, then the dog will not learn the sequence of steps that led to the reward. Essentially, the reward is not the result of a single step executed correctly. It's the result of a sequence. So how should credit be distributed?

This became known as the temporal credit assignment problem: when a reward arrives after a long sequence of actions, how do we determine which actions deserve credit?

Scientists needed a way to assign value not only to the final behavior but also to the intermediary actions that made the reward possible.

Richard Sutton's insight: temporal difference learning

To address this problem, Richard Sutton proposed a powerful idea that later became the foundation of reinforcement learning in AI.

Instead of reinforcing behaviors with actual rewards, what if you reinforced behaviors with predicted rewards?

In other words, an action should be rewarded not because it immediately produces a reward, but because it improves the system's prediction of a reward.

In the dog example, pressing the lever may not produce the chocolate directly. However, pressing the lever makes it possible for the dog to proceed to the next stage of the sequence. As a result, the probability of eventually obtaining the reward increases.

Any action along the path that creates a positive change in the prediction of future reward should themselves be reinforced.

This was a significant departure from the prevailing view that rather than learning only from rewards, intelligent systems should learn from improvements in their expectations of future rewards.

The actor-critic framework

To illustrate the point, Sutton alongside his colleague Andrew Barto and P. Anderson, proposed the modern computational actor-critic framework

In this framework, two components work together:

The actor selects actions.
The critic evaluates those actions and provides feedback.

When an action improves the prediction of future reward, the critic generates positive feedback. The actor then becomes more likely to choose the same actions in the future (positive reinforcement).

Conversely, when an action reduces the likelihood of obtaining a reward, the critic generates negative feedback, causing the actor to avoid the same actions (negative reinforcement).This framework provided a practical solution to the credit assignment problem and became one of the foundational ideas in reinforcement learning in machines.

But the question remained: Is that how the human brain actually works? At the time, nobody knew the answer.

Testing the theory: Dopamine and learning

While Sutton had hoped there was a connection between his idea and the brain, it was Peter Dayan, one of his colleagues, who found it. Peter Dayan and his colleague Read Montague were convinced that the brain implemented some form of temporal difference learning mechanism.

To investigate this, they turned their attention to Dopamine.

At the time, even now in popular conversations, dopamine was generally understood as a reward or pleasure molecule. Researchers knew that dopamine activity increased when animals received rewards, so it seemed natural to associate dopamine with pleasure.

However, experiments revealed something much more interesting. Read wolfram schultz’s experiments on macaque monkeys

Note: Again, not giving the example from the book my own example.

Consider a dog that hears a bell before receiving a piece of chocolate. Initially, the dog has not learned the association between the bell and the reward. When the dog receives the chocolate, dopamine neurons exhibit a strong burst of activity.

But after repeated trials, something unexpected happens.

The dopamine burst gradually shifts from the reward itself to the cue that predicts the reward. Eventually, the strongest dopamine response occurs when the bell rings, not when the chocolate arrives.

This led scientists to an important insight: dopamine is not merely a reward signal. Instead, it functions as a learning signal, encoding what researchers call a reward prediction error, the difference between expected and actual outcomes.

The dopamine responses align exactly with Sutton’s temporal difference learning signal. In other words, dopamine helps the brain learn whether the world is better or worse than expected.

Ok, the critic has been found. But who is the actor?

Dopamine as the Critic and the Basal Ganglia as the Actor

The basal ganglia is the seat of habits, which are automated motor responses or myelin sheaths (proteins+lipids) . Through repeated feedback from dopamine-based learning signals, it gradually strengthens behaviors that increase the likelihood of future rewards and weaken behaviors that do not.

So in this framework:

Dopamine systems act as the critic, evaluating outcomes and generating learning signals when expectations change.
The basal ganglia acts as the actor, selecting and reinforcing actions based on those signals.

Conclusion

There’s no larger purpose behind sharing this article. Like me, I am sure many of you get excited when you read something quite fascinating. The idea for this post stemmed from such a sense of fascination. With AI becoming such a central part of our everyday lives, I found it interesting to read a small part of the history behind its development.

We often think of neuroscience inspiring AI. In this case, however, AI research helped scientists formulate hypotheses about how the brain might solve a fundamental learning problem.

In other words, an idea developed to improve machines ended up revealing something about our brains. How exciting!!

Manifestation may help break bad habit patterns

Mon, 29 Jun 2026 13:07:12 +0530

I was listening to Raj Shamani's podcast with Dr. Vidita Vaidya, Neuroscientist, TIFR Mumbai, recently,
and one particular segment where Raj questions Dr Vaidya on social media addiction stood out for me.
I think it's somewhere around 1:37.25 to 1.45.

Raj asks a very relatable question: (paraphrasing)

"Why do so many people scroll endlessly through social media, regret it before going to bed, and then repeat the same behavior the next day? What happened to their willpower?"

Dr. Vaidya then goes on to explain that many addictive behaviors, whether social media scrolling, gambling, substance abuse, or even certain everyday compulsions, are driven by the brain's reward circuitry - same underlying circuits involved in food, sex etc - involving dopamine and the basal ganglia.

Raj subsequently asks how one can break away from such entrenched bad habit patterns. To that, her suggestion is that we learn new things, do the hard things, and delay gratification. In her view, novelty and hard work can help disrupt entrenched habit loops.

As I was listening, I couldn't help but be reminded of
Max Benett's "A Brief History of Intelligence", where Bennett discusses a wide array of topics from habit formation, goal-seeking behaviour, to exercising will-power and self-control. (BTW, it's a mind-blowing book that'll introduce you to an enormous range of concepts in
biology, evolution, brain science, and AI. I found it extremely fascinating)

In one of the chapters, Bennett discusses the role of dopamine in affecting reward-seeking behaviours and basal ganglia in solidifying these behaviours into habits. In essence, habits are semi-automated behavioural programs formed through reinforcement (repeated actions), mediated by the basal ganglia and dopamine-based reward signaling. Therefore, when a behaviour becomes a habit, it works on autopilot, in a way, requiring less effort/brain power. (That's when we call something a habit in the first place, right?)

Now, circling back to Raj's question to Dr. Vaidya,
when an addictive behaviour becomes a habit, how does one break away from it?

I mentioned Dr Vaidya's response above. However, Benett's discussion on self-control and willpower offers deeper insights into the above question. While he does not explicitly discuss addiction or breaking away from addictive behaviours, his explanation of the neuroscience of willpower and the role of the neocortex (a specialized region in mammals) and its relationship to basal ganglia and amygdala offers the clue.

Neocortex, basal ganglia, amygdala

The neocortex (prefrontal cortex in particular) functions as the brain's executive leader.
It regulates systems such as the basal ganglia, primarily responsible for habit formation, and the amygdala, primarily responsible for processing emotions.

This relationship is similar to that of a boss and his team in a corporate setting. Routine tasks are handled by employees without constant / minimal supervision (amygdala & basal ganglia). However, when an important strategic decision needs to be made, the boss steps in and
provides direction or has the power to change an established process. (Obviously, I am simplifying a lot; but the point is valid)

Now, carefully follow the line of thought. Benett argues that,
Attention, Working Memory, Self-control (often referred to as willpower), Planning are all different applications of the brain and are triggered by neocortical stimulation.The prefrontal cortex -particularly the agranular prefrontal cortex (aPFC), plays a key role in executing all of the above functions in mammals.

Here's the key point: Internal simulation - through imagination, visualization, or manifestation - is a form of neocortical stimulation (activates the aPFC).

What this proves is manifestation plays a key role in self-control and can influence behaviour.

When people talk about manifestation, they often focus on outcomes - manifesting a house, a car, a job, or moving abroad etc.
Many tend to think that if one simply visualizes these outcomes, the universe will somehow conspire to make them happen.
I am not ridiculing those who practice manifestation in the above manner. But it appears to me that the true power of manifestation lies in its ability to direct behaviour not in promising outcomes.

The question now is,
What does one manifest in the context of breaking away from an addictive behaviour? It's logical to manifest an "alternate behaviour", which is perhaps optimised for dopamine.

Here's a para from the book to support my hypothesis,

"How does the aPFC "control" behaviour? The idea presented here is that it doesn't control behaviour per se; it tries to convince the basal ganglia of the right choice by "vicariously showing" it that one choice is better and by filtering what information makes it to the basal ganglia. The aPFC controls behaviour not by telling but showing."
[emphasis added]
Pg 218 "How Mammals Control Themselves: Attention, Working Memory, and Self-control"

From a neuroscience perspective, repeated visualisation of a behaviour strengthens the neural representations associated with those behaviours, making it easier for the brain to execute them in real life.

How could this potentially work - my hypothesis

When the cue appears, eliciting a particular bad habit, manifesting the new behavior is the smart thing to do.
It does not indicate it'd be possible to execute the new behavior initially, given that the old one has solidified into a habit.
In fact, it's likely that one will fail many times at first. However, by repeatedly manifesting the new behavior, strengthening its neural representation, and gradually associating it with the cue, one can eventually enable the basal ganglia to express the new behavior.

I like to think of it as a vote. Every time the cue appears and you resist executing the old behavior, the brain negatively reinforces that old behavior. Every time you execute the new behavior, the brain positively reinforces it. As the new behavior accumulates enough "votes," it eventually surpasses a certain threshold, making it more likely to be expressed whenever that particular cue appears.

Given that the neocortex regulates systems such as the basal ganglia and the amygdala, I believe manifestation
can help people break away from unhealthy habit patterns and develop healthier ones.

As an aside, here's my picture, taken a few months ago, with Dr. Vidita Vaidya at IIT Madras, post her lecture on Neurobiology. After the session, she was kind enough to interact with me and a bunch of aspirig scientists from IIT Madras on various topics. She encouraged me
to start a podcast and even agreed to appear as a guest :)