Professor’s seminar on scientific literature paves the road to creative thinking

01/07/15 – BOSTON, MA. – Rebecca Shansky poses for a portrait on Jan. 7, 2015. Photo by Adam Glanzman/Northeastern University

Reading scientific literature—a must for any researcher—is akin to learning a new language. Jargon-filled sentences, often in the passive voice, can be indecipherable. Figures with cryptic, acronym-laden captions fly at you. Conclusions may sound reasonable, but how do you know the data is reliable or that confounding factors didn’t skew the results?

Rebecca Shansky, assistant professor in the Department of Psychology, offers a capstone course to help students start the translation process early. Called “Seminar in Biological Psychology: Animal Models of Mental Illness,” the course has a dual focus: Students learn—through reading more than 20 journal articles, giving presentations, and rigorous discussions—how to critically analyze academic papers while simultaneously gaining a solid footing in the literature exploring the ways animal research contributes to our understanding of the causes underlying psychiatric disorders.

“One of the really important things about learning to be a scientist is thinking critically and understanding that just because a paper is peer reviewed and published doesn’t mean that it’s without flaws,” says Shansky, who’s been teaching the seminar since 2013. “It’s important to take the time to understand the experimental design and the techniques and assumptions of the research because how the work was done can influence the interpretation of the findings.”

The road to creative thinking

An animal model of a mental illness could be a rat that exhibits signs of post-traumatic stress disorder after being exposed to a predatory cat and continual new cagemates for a designated period of time or one that exhibits symptoms of schizophrenia in adolescence because a lesion was introduced into a specific part of its brain at birth. Researchers perform experiments on such animals as an initial step to determine, for example, how a particular medication might affect the condition in humans.

Now when I read papers I don’t just accept the conclusion but investigate how convincing I think it is based on the data.
— Joseph Zaki, S’18

Using PowerPoint or Google Docs, students in Shansky’s class deconstruct and present published research that uses such animal models, reporting on every graph and figure illustrating the findings. Shansky then uses her knowledge of the field to push the students to dig deeper, think more critically.

“In behavioral neuroscience, researchers do something to an animal and it changes its behavior,” says Shansky. “In order to conclude why it’s changed its behavior, we need to rule out alternate explanations. So I might say to a student, ‘OK, the animal started walking faster. But does that mean it’s feeling more anxious or something else? What kinds of experiments could you do to zero in on the real driver?”

Joseph Zaki, S’18, who took the class this fall, can attest to the value of Shansky’s methods. “Now when I read papers I don’t just accept the conclusion but investigate how convincing I think it is based on the data,” he says. “For example, I might look to see if enough experiments were performed to rule out misinterpretations, or if enough subjects were in the study, say, 10 mice instead of just three, to justify the argument. Once you understand the literature, you can start thinking creatively as a scientist.”

Zaki, a behavioral neuroscience major and computer science minor, came into the class with some experience parsing journal articles. He had done two co-ops in neuroscience research labs, one at Harvard Medical School, where he explored the centers of smell in the mouse brain, and the other at Tufts Medical Center, where he plumbed the molecular mechanisms underlying fear, learning, and memory in mice. He credits Shansky’s class with opening his eyes to the many ways animal models can be used to understand not just illnesses such as addiction and schizophrenia but also coping mechanisms, including resilience to stress.

“The class really pushed me to the next level and helped me start to focus on what I want to study in grad school,” he says. For his final paper, he dove into the subject of spontaneous recovery in post-traumatic stress disorder—a little-studied phenomenon relevant to humans because it could reveal important information about relapse. “In my next co-op, at the Harvard Center for Brain Science, I want to explore that,” he says.

The tools to advance research

Shansky says that about half her students come in with experience reading journal articles, half without it. Ronan Talty, S’17, who has worked in Shansky’s Laboratory of Neuroanatomy and Behavior since his freshman year, had that experience from co-ops and research assistantships but says the class went further. “It has a broad scope,” he says. “It exposes you not just to research related to a particular lab but to the fundamental research in an entire subfield.”

This class gave us to the tools to tease apart the different factors that contribute to scientific outcomes.
— Ronan Talty, S’17

In addition to winning the 2015 RISE undergraduate poster award in the physical and life sciences category for research he did with Shansky, Talty has worked as a research assistant in two labs at the Icahn School of Medicine at Mount Sinai. The first examined how pathological brain activity contributes to generating seizures in epilepsy patients, and the second investigated the effects of charged particle irradiation on cognitive function in mice in an effort to learn more about the effects of irradiation during space travel.

Yet it was not until Shansky’s class that he had the opportunity to explore the progression of research techniques by looking at papers across the years, to compare and contrast papers that asked similar questions but came up with different conclusions, and to address the suitability of a particular animal model for the experiment at hand.

For example, he says, the compare and contrast approach forced students to investigate whether the researchers used different behavioral stressors, or rats versus mice, males versus females, or animals of different ages. The question about appropriate animal models invited students to both revise the experiment and suggest improvements to the model itself.

“To study a mouse exhibiting behavior modeling depression, a researcher might use the forced swim test,” says Talty, who is currently completing his honors thesis in the lab of Northeastern’s Greg Miller, associate professor in the Bouvé College of Health Sciences, and applying to combined medical and doctorate programs. “A shorter swim could indicate depression. But it could also mean that the mouse just realized that once it stopped swimming it would be taken out of the water.”

“This class,” he says, “gave us to the tools to tease apart the different factors that contribute to scientific outcomes.”