The Genetic Link Between Creativity and Psychiatric Disease


The biological sciences are in a golden era: the number of advanced technological tools available coupled with innovations in experimental design has led to an unprecedented and accelerating surge in knowledge (at least as far as the number of papers published is concerned). For the first time in history, we are beginning to ask questions in biology that were previously unanswerable.

No field demonstrates this better than genetics, the study of DNA and our genes. With the advent of high-throughput DNA sequencing, genetic information can be acquired literally from thousands of individuals and even more remarkably, can be analyzed in a meaningful way. Genomics, or the study of the complete set of an organism’s DNA or its genome, directly applies these advances to probe answers to questions that are literally thousands of years old.

A recent study, a collaborative effort from scientists in Iceland, the Netherlands, Sweden, the UK, and the US, is an example of power of genomics and to answer these elusive questions.

Power eet al. Nat. Neursci. 2015. Title

The scientists posed an intriguing question: if you are at risk for a psychiatric disorder, are you more likely to be creative? Is there a link between madness and creativity?

Self-portrait with bandaged ear. Vincent van Gogh, 1889. (
Self-portrait with bandaged ear. Vincent van Gogh, 1889. (

Aristotle himself once said, “no great genius was without a mixture of insanity” and indeed, the “mad genius” archetype has long pervaded our collective consciousness. But Vincent Van Gogh cutting off his own ear or Beethoven’s erratic fits of rage are compelling stories but can hardly be considered empirical, scientific evidence.

But numerous studies have provided some evidence that suggests a correlation between psychiatric disorders and creativity but never before has an analysis of this magnitude been performed.

Genome-wide association studies (GWAS) take advantage of not only the plethora of human DNA sequencing data but also the computational power to compare it all. Quite literally, the DNA of thousands of individuals is lined up and, using advance computer algorithms, is compared. This comparison helps to reveal if specific changes in DNA, or genetic variants, are more common in individuals with a certain trait. This analysis is especially useful in identifying genetic variants that may be responsible for highly complex diseases that may not be caused by only a single gene or single genetic variant, but are polygenic, or caused by many different genetic variants. Psychiatric diseases are polygenic, thus GWAS is useful in revealing important genetic information about them.

This video features Francis Collins, the former head of the Human Genome Project and current director of the National Institutes of Health (NIH), explaining GWAS studies. The video is 5 years old but the concept is still the same (there’s not many GWAS videos meant for a lay audience).

The authors used data from two huge analyses that previously performed GWAS on individuals with either bipolar disorder or schizophrenia compared to normal controls. Using these prior studies, the author’s generated a polygenic risk score for bipolar disorder and for schizophrenia. This means that based on these enormous data sets, they were able to identify genetic variants that would predict if a normal individual is more likely to develop bipolar disorder or schizophrenia. The author’s then tested their polygenic risk scores on 86,292 individuals from the general population of Iceland and success! The polygenic risk scores did associate with the occurrence of bipolar disorder or schizophrenia.

Next, the scientists tested for an association between the polygenic risk scores and creativity. Of course, creativity is a difficult thing to define scientifically. The authors explain, “a creative person is most often considered one who take novel approaches requiring cognitive processes that are different from prevailing modes of thought.” Translation: they define creativity as someone who often thinks outside the box.

In order to measure creativity, the authors defined creative individuals as “belonging to the national artistic societies of actors, dancers, musicians, and visual artists, and writers.”

The scientists found that the polygenic risk scores for bipolar disorder and schizophrenia each separately associated with creativity while five other types of professions were not associated with the risk scores. An individual at risk for bipolar disorder or schizophrenia is more likely to be in creative profession than someone in a non-creative profession.

 The authors then compared a number of other analyses to see if this effect was due to other factors such as number of years in school or having a university degree but this did not alter the associations with being in a creative field.

Finally, the same type of analysis was done with two other data sets: 18,452 individuals from the Netherlands and 8,893 individuals from Sweden. Creativity was assessed slightly differently. Once again creative profession was used but also data from a Creative Achievement Questionnaire (CAQ), which reported achievements in the creative fields described above, was available for a subset of the individuals.

Once again, the polygenic risk scores associated with being in a creative profession to a similar degree as the Icelandic data set; a similar association was found with the CAQ score.

The authors conclude that the risk for a psychiatric disorder is associated with creativity, which provides concrete scientific evidence for Aristotle’s observation all those years ago.

However, future analyses will have to broaden the definition of creativity beyond just narrowly defined “creative” professions. For example, the design of scientific experiments involves a great deal of creativity but is not considered a creative profession and is therefore not included in these analyses, and a similar argument could be made with other professions. Also, no information about which genetic variants are involved or what their function is was discussed.

Nevertheless, this exciting data is an example of the power that huge genomic data sets can have in answering fascinating questions about the genetic basis of human behavior and complex traits.

For further discussion, read the News and Views article, a scientific discussion of the paper, which talks about potential evolutionary mechanisms to explain these associations.


The Formation of New Memories in the Human Brain

Image of the structure of the mouse Hippocampus (Image courtesy of
Image of the structure of the mouse Hippocampus (Image courtesy of

How are new memories created?

This is a fascinating question in neuroscience and at the very core of what makes us human. After all, our entire concept of ourselves is defined by our memories and without them, are we even ourselves? This is a pretty lofty philosophical discussion… but today we’re only interested in the neuroscience of memory.

In specific, what happens to individual neurons in the human brain when a new memory is created and recalled?

Researchers at the University of California-Los Angeles performed a study in humans that has shed some light on this important question. Published recently in the journal Neuron, the novelty of the study involved recording how many times a neuron would fire during a specially designed memory test. In other words, the scientists were able to monitor what happened to individual neurons in a human being as a new memory was being created!

Title Ison et al. 2015

This article is open access (able to downloaded and distributed for free). The article can be found here or download the pdf.

Before I go into what the researchers found, let’s see how it was done.

The subjects in the study were patients being treated for epilepsy. As part of their clinical diagnosis, they had been implanted with an electrode, a tool used to measure neuronal activity or in other words, the electrode measures how often a neuron fires. The fact these patients already had an electrode inserted into the brain for clinical reasons made it convenient for the researchers to conduct this study.

Left Temporal Lobe (
Left Temporal Lobe (

The brain region in which the electrode was implanted is called the medial temporal lobe (MTL). The image to the right is of the left human temporal lobe. The medial region of the temporal lobe is located more towards the center of the brain.

Human Hippocampus (
Human Hippocampus (

One specific region of the MTL, the hippocampus, is believed to be the primary brain region where memories are “stored”. Specifically, previous studies in animals and humans have suggested that the MTL and hippocampus are very important to encoding episodic memory. Episodic memory involves memories about specific events or places. In this study, the example of episodic memory being used is remembering seeing a person at a particular place. Another example: the game Simon™ can be considered a test of your brain’s ability to rapidly create and recall short-term episodic memories!

Simon game memory

*Note: Episodic memory is considered one of the main bifurcations of declarative memory, or memories that can be consciously recalled. The other type of declarative memory is semantic memory, which are memories of non-physical/tangible things, like facts.

To test the episodic memory of remembering a person at a particular place, images were presented to the patients while the neurons were being recorded. There were 5 different tasks (all completed within 25-30min). See Figure 1 below from the paper.

Figure 1: Experimental Design
Figure 1: Experimental Design

First, a pre-screening was done in which the patients was shown many random images of people and places. The activity of multiple neurons was recorded and the data was quickly analyzed then 3-8 pairs of images were compiled. In each pair, 1 image was “preferred” or “P” image, meaning the neurons being recorded fired when the “P” image was shown. The second image was “non-preferred” or “NP” image, meaning the neurons did not respond to it when it was shown.

The first task is the “Screening” test. Each “P” and “NP” image was shown individually and the neurons response to each was recorded. As you would expect, the neuron would fire heavily to the “P” image and not very much to the “NP” image.

The second task was the “learning task” in which a composite image of each of the “P” and “NP” image pairs was made. The person in the “P” image was digitally extracted and placed in front of the landmark in the “NP” image. After the composite images were shown, the individual images were shown again.

For example, in one image pair for one patient, the “P” image was a member of the patient’s family while the “NP” image was the Eiffel Tower (for this example, see Figure 2). The composite image in the “learning” task was the family member in front of the Eiffel Tower. Another example of a “P” image was Clint Eastwood and the “NP” image was the Hollywood sign. The composite image would therefore be Clint Eastwood in front of the Hollywood sign. (However, in some image pairs the “P” image was a place and “NP” image was a person).

The third task was “assessing learning”. The image of just the person in the composite image was shown and the patient had to pick out the correct landmark he/she was paired with. For example, the picture of the family member was shown and the patient would have to pick out the Eiffel Tower image.

The fourth task was the “recall” task. The landmark image was shown and the patient had to remember and say the person it was paired with. For example, the Eiffel Tower was shown and the patient had to say the family member’s name.

Finally, the fifth task was a “re-screening” in which each individual image was shown again so the neuron’s activity could be compared to the Task 1, pre-learning.

The activity of multiple neurons were recorded for each image for each of the tasks. The data was then analyzed in number of different ways and the activity of different neurons was reported.

And what was found?

Figure 2: Response of Neruons in the Hippocampus from One Patient
Figure 2: Response of Neurons in the Hippocampus from a Patient

Let’s go back to the family member/Eiffel tower example. The researchers were able to show that a neuron in the hippocampus responded heavily to the picture of the family member (“P” image) but not to the Eiffel Tower (“NP” image). After showing the composite image, the neuron now responded to the Eiffel Tower too in addition to the family member! (The neuron also fired a comparable amount to the individual family member image as the composite image).

As you can see in Figure 2, each little red or blue line indicates when a neuron fired. For example, in Task 1 you can clearly see more firing (more lines) to the “P” image than the “NP” image. You can see that after Task 2, the neuron responds to either the “P” or “NP” image (especially obvious in the Task 5). The middle graph indicates the firing rate of the neurons to the “NP” image and it clearly shows increased firing rate of the neuron after learning (AL) compared to before learning (BL). It may look small, but the scientists calculated a 230% increase in firing rate of the neuron to “NP” image after the learning/memory task took place!

What does this mean? It means that a new episodic memory has been created and a single neuron is now firing in a new pattern in order to help encode the new memory!

This was confirmed the other way around too. In another patient, the “P” neuron was the White House and the “NP” image was beach volleyball player Kerry Walsh. The neuron that was being recorded fired a lot when the image of the White House was shown but not so much for the Kerri Walsh image. Then the composite image was shown and the learning/recall tasks were performed. The neuron was shown to fire to both the White House image AND the Kerry Walsh image! The neuron was responding to the new association memory that was created!

Keep in mind these are just two examples. The scientists actually recorded from ~600 neurons in several different brain regions besides the hippocampus but they only used about 50 of them that responded to visual presentation of the “P” image, either a person or a landmark (the identification of visually responsive neurons was crucial part of the experiment). Remarkably, when the firing rates of all these neurons was averaged before and after the memory/learning tasks, a similar finding to the above examples was found: the neuron now responded to the “NP” image after the composite was shown!

Many other statistical analyses of the data was done to prove this was not just a fluke of one or two neurons but was consistent observation amongst all the neurons studied but I won’t go into those details now.

But what’s going on here? Are the neurons that respond to the “P” stimulus now directly responding to the “NP” image or is more indirect, some other neuron is responding to the “NP” which in turn signals to the “P” neuron to increase in firing? The authors performed some interesting analyses that both of these mechanisms may apply but for different neurons.

Finally, were all the recorded neurons that were engaged in encoding the new episodic memory located in the hippocampus? The answer is no. Responsive neurons were identified in several brain regions besides the hippocampus including the entorhinal cortex and the amygdala. But most of the responsive cells were located within the parahippocampal cortex, a region of the cortex that surrounds the hippocampus, thus not surprising it is very involved in encoding a new memory.

In conclusion, the scientists were able to observe for the first time the creation of a new memory in the human brain at the level of a single neuron. This is an important development but such a detailed analysis has never before been done in humans and, most importantly, in real time. Meaning, the experiment was able to observe the actual inception of a new memory at the neuronal level.

However, one major limitation is that the activity of these neurons were not studied in the long term so it’s unknown if the rapid change in activity is a short-term response to the association of the two images or if it really represents a long-term memory. The authors acknowledge this limitation but the problem is really in the difficulty of doing such studies in humans. It’s not really ethical to leave an electrode in someone’s brain just so that you can test them every week!

But what does all of this mean? The authors do suggest that the work may help to resolve a debate that has been going in on the psychology field since the 40s. Do associations form gradually or rapidly? These results strongly suggest new neurons rapidly respond to encode the new memory formation.

But how will these results shape the neuroscience of memory? The answer is I don’t know and no one does. Thus is the rich tapestry of neuroscience, another thread weaved by the continuing work of scientists all over the world  in order to understand what it is that makes us human: our brains.

Morphine and Oxycodone Affect the Brain Differently

(Neurons. Image from Ana Milosevic, Rockefeller University)
(Neurons. Image from Ana Milosevic, Rockefeller University)

Why are some drugs of abuse more addictive than others?

 This is a central question to the addiction field yet it remains largely a mystery. All drugs of abuse have a similar effect on the brain: they all result in increased amounts of the neurotransmitter dopamine (DA) in an important brain region called the mesolimbic pathway (also known as the reward pathway). One of the core components of this pathway is the ventral tegmental area (VTA), which contains many neurons that make and release DA. VTA neurons communicate with neurons in the nucleus accumbens (NAc). This means that the axons of VTA neurons project to and synapse on NAc neurons. When VTA neurons are stimulated, they release DA onto the NAc, and this is a core component of how the brain perceives that something is pleasurable or “feels good.” Many types of pleasurable stimuli (food, sex, drugs, etc.) cause DA to be released from the VTA onto the NAc (See the yellow box in the diagram below). In fact, all drugs of abuse cause this release of DA from VTA neurons onto NAc neurons.

*Important note: many other brain regions are involved in how the brain perceives the pleasurable feelings of drugs besides the VTA and NAc, but these regions represent the core of the pathway.

"Dopamineseratonin". Licensed under Public Domain via Wikipedia.
“Dopamineseratonin”. Licensed under Public Domain via Wikipedia.

Check out these videos for a more detailed discussion of the mesolimbic pathway.

But if all drugs of abuse cause DA release, then why do different drugs make you feel differently? This is a very complicated question but one component of the answer is that different drugs have different mechanisms and dynamics of DA release.

For the opioid drugs like heroin, morphine, and oxycodone, they are able to bind to a special molecule called the Mu Opioid Receptor (MOPR). This action on the MOPR results in an indirect activation of DA neurons in the VTA and a release of DA in the NAc. While all opioid drugs reduce the feeling of pain and induce a pleasurable feeling, they have slightly different properties at the MOPR.

The different properties of the opioids may be a reason why some are more abused than others. For example, a number of studies have suggested that oxycodone may have greater abuse potential than morphine. This means that oxycodone is more likely to be abused morphine.

But do the different properties of morphine and oxycodone on the MOPR affect DA release and is this important to why oxycodone is more likely to be abused than morphine?

Vander Weele et al. 2015 titleThis is the question that scientists at the University of Michigan sought to address. Using several different sophisticated techniques, the scientists looked at differences in DA release in the NAc caused by morphine and oxycodone, two common opioid drugs.

Rats were injected with either morphine or oxycodone and then DA release was measured using either fast-scan cyclic voltammetry or microdialysis. I’ve discussed microdialysis in a previous post but in brief, it involves drawing fluid from a particular brain region at different time points in an experiment and then measuring the neurotransmitters present (using advanced chemistry tools that I won’t explain here).

Voltammetry is a more technically complicated technique. In brief, it uses electrodes to measure sensitive voltage changes. Since a molecule has specific electrochemical properties, these voltage changes can be related back to a specific molecule, such as the neurotransmitter DA as in this study. Voltammetry may even allow greater temporal resolution (easier to detect very precise changes at very short time frames, like seconds), which may make it more accurate than microdialysis (which can only measure neurotransmitter release on the scale of minutes).

Because each technology has its own limitations and potential problems, the authors used both of these techniques to show that they are observing the same changes regardless of the technology being used. Showing the same observation multiple times but in different ways is a common practice in scientific papers: it increases your confidence that your experiment is actually working and what you are observing is real and not just some random fluke.

The authors administered a single dose of either morphine or oxycodone to rats and then measured the DA release in the NAc as described above. What they found were very different patterns!

Morphine resulted in a rapid increase in DA (less than 30 seconds) but by 60 seconds had returned to normal. In contrast, oxycodone took longer to rise (about 20-30 sec before a significant increase was detected) but remained high for the entire 2 minutes that it was measured. The difference in DA release caused by morphine and oxycodone is striking!

Many other changes were observed such as differences in DA release in different sub-regions of the NAc, different effects on phasic release of DA (DA is often released in bursts), and differences in the other neurotransmitters such as GABA (morphine caused an increase in GABA release too while oxycodone did not). I won’t discuss these details here but check out the paper for more details.

Of course, do these differences in DA release explain why oxycodone is more often abused than morphine? Unfortunately no, there are many other factors (for example, oxycodone is more widely available than morphine) to consider. Nevertheless, this is some intriguing neuroscientific evidence that adds one more piece to the addiction puzzle.