The first time I saw the term "limbic system" in a textbook was in 1997, during my sophomore year neuroanatomy course. That third edition of Kandel's Principles of Neural Science—I still have it on my office bookshelf, the spine already cracked.
Back then, our understanding of the limbic system was simple. The hippocampus handles memory, the amygdala handles fear, the hypothalamus handles endocrine function. The professor drew a circle on the blackboard, enclosed these structures within it, and said, "This is your emotional center."
It took me more than twenty years to realize how problematic that circle was.
The brain's complexity defies simple categorical boundaries
The concept of the limbic system itself is a historical artifact. When Paul MacLean proposed this term in 1952, he was actually promoting his triune brain theory—the reptilian brain, paleomammalian brain, and neomammalian brain. Nobody takes that theory seriously anymore, but the term "limbic system" survived.
2008I wrote a review article attempting to sort out the definitional problems of the limbic system. I submitted it to Neuroscience & Biobehavioral Reviews. The reviewer's comment was: this problem is too old, nobody cares anymore. He was right. Textbooks still use the term, undergraduates still memorize this concept, clinicians still use "limbic" to describe certain symptoms.
We've stuffed too many things into the basket called "emotion."
— Joseph LeDoux, SfN Annual Meeting, 2014
2014I attended the SfN annual meeting and encountered Joseph LeDoux at a small symposium. He was writing his book Anxious at the time. We talked for about twenty minutes, mostly with me listening. He said something I still remember to this day: "We've stuffed too many things into the basket called 'emotion.'"
That statement changed my research direction going forward.
LeDoux published an article in 2015 with a chapter titled "The Limbic System Myth" in his book Anxious: Using the Brain to Understand and Treat Fear and Anxiety. When I saw it at the time, I thought: finally, someone said it.
When I teach Introduction to Neuroscience now, I still cover the limbic system. There's no way around it. Students need to pass licensing exams, and the exams test this. I spend an entire class period telling them about the problems with this concept.
There are several problems.
Ask ten neuroscientists which structures the limbic system includes, and you'll get ten different answers. Some count the orbitofrontal cortex, some don't. Some count the insula, some don't. Some count the nucleus accumbens, others object. A review in the 1990s surveyed different textbooks and found that definitions of the limbic system involved anywhere from 5 to 23 brain regions.
The hippocampus is indeed related to memory, but it's also related to spatial navigation. The amygdala is indeed related to fear, but it's also related to reward learning and social cognition. Drawing a circle around a group of structures and saying they handle emotion—that statement is too crude.
MacLean's triune brain theory implies the limbic system is primitive, something we share with reptiles and early mammals. In reality, mammalian structures differ greatly from their reptilian counterparts. Homology doesn't equal functional equivalence. Georg Striedter explained this problem clearly in his 2005 book Principles of Brain Evolution.
Modern neuroimaging challenges traditional categorical thinking
2019One of my students completed a project. She used resting-state fMRI to examine the functional connectivity strength between various structures of the so-called limbic system. The results showed that connectivity between these structures wasn't stronger than their connectivity with non-limbic structures. From a functional connectivity perspective, the limbic system doesn't hold together as a system.
During her defense, a committee member asked: "So what should we call these structures?" She said: "Just call them by their names. The hippocampus is the hippocampus, the amygdala is the amygdala. We don't need a superordinate concept to bundle them together."
I agreed with her. I also know this change won't happen quickly.
The Inertia of Terminology
Scientific concepts have their own life cycles. Some concepts are problematic from birth but survive because they're useful. The limbic system is one of these. It gives clinicians a label, gives textbook authors a way to organize chapters, gives students a framework.
Useful. Not accurate.
Now, every time I see the term "limbic system" in a paper, I look more carefully, wanting to know what the author actually means. Most of the time they mean the amygdala. Occasionally they mean the hippocampus. Sometimes they mean the hypothalamus or cingulate cortex. Almost no one actually means that circle MacLean drew in 1952.
The gap between textbook concepts and research reality persists
2020Someone initiated a petition suggesting that academic publications use more precise terminology instead of "limbic system." I signed it. The petition had about two hundred signatures, including several well-known people in the field. And then nothing happened.
Journals didn't change their submission guidelines. Textbooks didn't change their chapter organization. Exams didn't change their syllabi.
I understand. The cost of changing terminology is high. Textbooks need to be rewritten, teachers need to be retrained, exams need to be redesigned. All of this requires money and time. Most people who use "limbic system" know the concept is problematic; they're just using a conventional shorthand.
A conventional shorthand, used long enough, becomes a default thinking framework. I've seen too many clinical case reports attributing patients' symptoms to limbic system dysfunction. This kind of attribution sounds like an explanation but actually explains nothing. It just links a phenomenon that needs explaining to a poorly defined concept.
Handling It in Teaching
Last year an undergraduate came to see me, saying she wanted to do research on emotion regulation. She asked me whether limbic system overactivation is the cause of anxiety disorders.
I spent about forty minutes explaining the problems with the question itself. The limbic system isn't something whose activation level can be measured. "Overactivation" assumes there's a baseline of normal activation, which varies by person, by situation, by measurement method. Anxiety disorder is a clinical diagnosis, not a neural mechanism.
She eventually chose amygdala-anterior cingulate cortex functional connectivity as her entry point. Her paper turned out quite good.
The classroom: where concepts are transmitted
And where they must be carefully unpacked
I have conversations like this several times every semester. Students come with concepts from textbooks, and I have to spend time taking these concepts apart, helping them build a framework that better corresponds to actual research operations.
Sometimes I feel like I'm doing repetitive work. Then I think, maybe this is part of teaching. The function of textbooks is to let beginners quickly build a rough cognitive structure. My job is to tell them where that structure needs correction.
Some Records
I looked through my own notes. From 2003 to now, I've probably written more than three hundred pages of material about the problems with the limbic system concept. Most of it unpublished. Some are lecture notes for students, some are responses to reviewers (several times reviewers insisted I use the term "limbic system," and I wrote lengthy replies explaining why I don't), some are just my own thinking records.
2006I wrote a passage:
After writing it, I felt it was too pessimistic and deleted it. Now I think I should have kept it.
Last year I organized all my notes about the limbic system and put them on my personal website. I didn't promote it at all, just left it there. By now there have been about three thousand visits. I don't know who the visitors are. Maybe students who found it while doing homework. Maybe colleagues who came across it while preparing course materials. Maybe patients' family members who found it while trying to understand a diagnosis report.
Whether these materials are useful to them, I don't know either. They're just there, available when someone needs them.
