Behind the Scenes of Behavioural Experiments


Studying the brain to understand human behaviour involves elaborate experimental set-ups. Researchers at IISc describe how they design these experiments

A setup to track the participant’s eye gaze while restricting head movement (Photo: KG Haridasan)


Ever walked into a room and forgot why you went in there? Or struggled to keep your New Year’s resolutions? We have all shown such behaviours at some point or the other. The field of psychology tries to understand this: why we behave the way we do.

While psychology deals with the mind and behaviour, the study of the underlying biological components and processes comes under the realm of neuroscience. Though it borrows some experimental techniques from psychology, neuroscience investigates the activity of brain cells (called neurons) when we display certain behaviours.

Scientists have known for decades that neurons are specialised for transferring information to other neurons, as well as to other cells of the body. One might think that, if we know what a cell does, how hard would it be to figure out the function of an organ that is made up of those cells? When it comes to the brain, it is especially hard. And it gets harder when you move from the level of neurons to behaviour, which arises from a complex interplay of different neuronal networks in the brain. However, scientists have found a way around this problem: through behavioural experiments, which form the basis for much of neuroscience research.


What is a behavioural experiment?

A behavioural experiment is one of the tools that both neuroscientists and psychologists use to study behaviour. “There is a task that you tell or train the participant to do. Typically, you show them a stimulus and you get a response,” explains SP Arun, an Associate Professor at IISc’s Centre for Neuroscience (CNS).

In fact, behavioural experiments are not very different from what most of us do in our daily lives. For example, if you have ever tried to guess your dad’s mood from the pitch of his voice – especially when you need him to sign your report card – by asking him a trivial question like how his day was, then you have successfully conducted a behavioural experiment.


If we know what a cell does, how hard would it be to figure out the function of an organ that is made up of those cells? When it comes to the brain, it is especially hard


Neuroscientists, however, study much more than someone’s mood. They also study how the brain perceives the five senses (sight, sound, taste, touch and smell), higher level cognitive processes like attention and emotion, and motor control, or the ability of the brain to control our voluntary movements.

But how do they study these behaviours? What goes into designing a successful behavioural experiment?


Chalk out a plan

Just like an architect begins with a basic blueprint before building a house, a behavioural scientist starts by designing an experiment. A lot of thought goes into this part.

Let us say we want to study how long it takes for people to see the colour difference between two objects. How would one go about designing an experiment to study this? We could show people images of apples and oranges and ask if the two images are the same or different. By analysing their response time, we could get a fair estimate of how long people take to perceive colour difference.

But there is a major flaw in this theory.

When a person sees apples and oranges, how do we know that they are seeing the difference only based on colour? It could be based on size difference. Or, on the categorical difference – that they are two different fruits. Thus, if we want this experiment of ours to truly test for colour perception, we should show two identical circles, similar in every possible way other than their colour.

“Usually you’re trying to make sure that the participant is using the clues or the criteria that you want them to use. Just because you ask them to use a strategy, it doesn’t guarantee that they are going to use it. So, to really make sure that your participant is solving the task in the way that you intended, you should have a lot of controls,” explains Arun.

The controls that Arun talks about are the variables that might differ, like the size and category of the images in our experiment. These major variables are checked off during trial runs or ‘piloting’, when the experiment is done with a small set of people. Other trivial variables like the brightness of the screen or in the room are accounted for, and maintained constant across trials. The experiment is perfected through careful repeated trials to ensure that all possible gaps are accounted for.


Recruit participants

Once a perfect behavioural experiment is designed, the next step is to seek out participants. Posters and emails are circulated, mostly within the university, calling for willing participants, who are usually paid a fee for their participation.

Sometimes, behavioural experiments are complemented by modern techniques like functional Magnetic Resonance Imaging (fMRI), Electroencephalography (EEG), Magnetoencephalography (MEG) or Transcranial Magnetic Stimulation (TMS). These additional methods help researchers observe and record the activity of the brain while a person is doing the experiment. This allows the researcher to determine which part(s) of the brain is responsible for a particular behaviour.

The JN Tata MRI facility established at IISc with support from the Tata Trusts (Photo courtesy: CNS, IISc)


It is not always easy to recruit participants. For instance, fMRI experiments need people who do not have any kind of metal implants and are not claustrophobic. Some behavioural experiments recruit people who are not under any kind of medications, to eliminate the effects of drug interference. Other times, the experiment requires participants with specific neurological disorders in order to study how their behaviour differs from those without the disorder. This helps researchers come up with effective diagnostic tools for that disorder.

In a university, participants are mostly recruited from among the students. In a 2010 study, scientists criticised this practice by arguing how most research revolves around WEIRD – Westernised, Educated, Industrialised, Rich and Democratic – people. The study cautions against making sweeping claims based on a small and specific sample of participants.

But there is a way to reduce this bias. “One good thing is that we go outside the Institute and recruit people who are not just students, and are naive to the research,” explains Lakshman Chakrav, a Research Assistant in the Emotion and Cognition Lab at CNS. “This is to have generalisability to the research that we do. We can do a study [with a limited number of participants] and publish it, but two years from now, if no one can reproduce it, there’s no point in wasting public money.”

Once the participants are recruited, they are asked to fully understand the study before giving their written consent. A core tenet of human ethics policies is that participants’ refusal should not have any adverse consequences for them – this can be important in hospital settings where they receive care, or in university settings where they might receive grades.


Experiment begins

Different neuroscience questions call for different kinds of experiments. A lab that studies how our vision works, for example, would do something similar to the colour perception experiment. Whereas a lab that studies how our brain controls our muscles and joints for movement would require its participants to move their eyes, hands or head during the experiment.

Various metrics are used to measure the outcome of the experiment. For example, in the colour experiment, the time taken for the participant to respond would be a good metric that would answer our question on colour
perception. Similarly, a lab that studies eye movement would use eye trackers to see how fast our eyes move or which way they move. Software tools enable these metrics to be recorded by computers for further analysis.


Experience of an experimenter

From a participant’s point of view, an experiment may just be a fun way to spend time, or perhaps an exciting game that also helps them earn money. But for the experimenter, it may lead to some nail-biting moments, as they can see how each participant’s response becomes a data point in support of, or against, their hypothesis.

These trials can also test the experimenter’s patience. Surprisingly, the most common problem they face during an experiment is that participants sometimes doze off!

“I can usually see that after the first 100 trials, people get tired and their eyes begin to droop. Many people end up sleeping with their head at the chin rest,’’ says Kavya Rajendran, a PhD student at the Cognition Lab in CNS. Jana, too, recalls a particular participant who came in for the task and went on to sleep for two uninterrupted hours. fMRI experiments have it worse, as many people find the dull monotonous sound of the magnet soporific, and end up falling asleep comfortably inside the MRI chamber, rendering the data that was collected until then completely useless. Who knew that the fringe benefits of research life would include acquiring such essential life skills as patience and fortitude?

A behavioural experiment, apart from its academic significance, also helps make new acquaintances. Experimenters often meet interesting people, and these acquaintances lead to stimulating and interesting discussions.

“You get to meet so many people. It’s nice to get to know a little about the story behind a person. I learn about people’s families, why they came into this field, their pets and the books they read. I sometimes get good recommendations too. The interaction also helps people [participants] to become more engaged and motivated,” Priyanka Gupta, another PhD student at the Cognition Lab, says excitedly.

An EEG experiment underway in Supratim Ray’s lab at CNS, IISc (Photo: Atul Gopal)


Experiments in the time of COVID-19

While behavioural experiments may lead to invigorating conversations and connections – and hopefully important scientific insights – they also involve close human contact. But that isn’t possible during the current pandemic. Many labs, including at IISc, were either designing experiments, recruiting participants or in the middle of conducting experiments when the pandemic hit, forcing them to shut down their research.

The experiments conducted at the Cognition Lab, for example, focus specifically on attention, and were significantly impacted by the pandemic-related lockdown. Priyanka says, “It was all of a sudden, and our contingency plan was to work with the data we had already collected till then. We closed in March and I haven’t run a single human experiment since then.”

However, as the world is waiting to recover from this viral onslaught, cognitive research too is slowly waking up from its pandemic-induced slumber. With labs slowly re-opening, work is resuming steadily. New rules are in place, ensuring reduced risk of infection while conducting behavioural experiments. These include using ultraviolet lamps that kill viruses and disinfection sprays, ventilating rooms before and after each participant does an experiment, and wearing masks and gloves.


Learn, adapt and overcome

Some of Priyanka’s colleagues have also turned to other possible ways to plough through during the pandemic. As classrooms go live and virtual meetings replace physical conferences, unsurprisingly, even behavioural experiments are going online. Though sophisticated tasks using techniques like fMRI cannot be done online, researchers are adapting to the new normal by designing simple experiments that can be done from the comfort of the participant’s home.


Researchers are adapting to the new normal by designing simple experiments that can be done from the comfort of the participant’s home


But there is a catch to this as well. Arun explains, “It is a much more controlled setup because the same set of participants use the same computers. But now with experiments going online, there is this extra variability – whether they are using a big or a small laptop; sitting in a bright or a dark room.” It’s not all bad, he adds. “We have run some experiments online and are happy to find that the results are very similar to what we get in a lab-based set-up.” Another limitation is that only a small percentage of those that can afford to access the Internet can participate.

When asked if he thinks online experiments would be prevalent post the pandemic, Arun says yes. “There is certainly a range of experiments one can do online. Doing it online means you can send it to 100 people and they would do it from the safety of their homes. There is potential to collect data really, really quickly. In [the] online mode, one can also reach out to a more diverse set of people. I think it certainly allows us to do more interesting things.”


Rohini Murugan, a science writing intern in the Office of Communications (OoC) at IISc, is an Integrated PhD student in the Centre for Neuroscience

Post Author: