Why did curiosity kill the cat? Brain reward systems engaged in learning curious information

People, like cats, are curiosity-driven creatures that manifest unique inclinations towards certain stimuli. Curiosity´s unavoidability has been immortalized in the proverb of the cat. Could we be so extremely pulled by this impulse to the level of endangering ourselves for the mere sake of its satisfaction?

Why curiosity did kill the cat?  Kitten in a fishbowl | | Credit: wwm.gatitosgraciosos.com
Figure 1. Why did curiosity kill the cat?Kitten in a fishbowl || Credit: wwm.gatitosgraciosos.com

The flip side of the coin is an intriguing phenomenon: how curiosity potentiates learning? We all know that learning new information is generally difficult and requires tedious repetition, unless the topic really touches our fascination. Which is the neuronal basis for such effect of curiosity upon learning?

Knowing the neurobiological substrates of human learning generates great expectation in the educative community. It is thought that such information might help design more efficient didactic programs 1. The importance of curiosity or intrinsic motivation in the learning process has been already stressed in the past by classic theorists of pedagogy. In regards to brain mechanisms, neuronal processes by which external stimuli modulate learning, either extrinsic negative (punishment) or positive (reward) reinforcers, have been explored in much more detail than those underlying the effects of intrinsic factors or motivation, which remain mostly enigmatic.

Now, a team of researchers at the University of California have revealed some interesting clues into the brain substrates of curiosity and its influences on memory in a recent paper just published in advance online (October, 2nd) in Neuron 2. Researchers hypothesized that curiosity should manifest in the brain as an activation of reward circuits, those same that are elicited by pleasant external stimuli such as sex or food and that are known to potentiate learning. Plainly, whatever touches our deeper curiosity, researchers speculated, it would do so because it is rewarding, and that´s why we may learn it better.

Curiosity and reward

To address their hypothesis, researchers used functional Magnetic Resonance Imaging (fMRI) to compare the brain activity of adult subjects while they were learning information for which they felt a high or low curiosity. The material to be learnt was questions and answers from the famous Trivia game. In a preliminary phase, researchers selected what questions participants could not answer, and asked participants to indicate how curiosity-evoking those questions were. Later, researchers performed a brain scan while participants observed the subset of unknown Trivia questions eliciting varied levels of curiosity.

First, researchers discovered that the level of curiosity correlated well with the level of activity in two brain areas that process reward (one was a region in the mid part of the brain containing the Substantia Nigra and the Ventral Tegmental Area –SN/VTA-; the second region was the Nucleus Accumbens, located more anteriorly in the brain). Thus, these brain structures were more active if the participant found the Trivia question very interesting, and less active if the question was boring. The activation happened just after participants saw the questions, before the answer appeared. Then, such brain activity paralleled a mental state of high curiosity while anticipating the answer, and not the satisfaction of the curiosity itself.

20-40 minutes after the brain scan session, participants did a memory test. Confirming previous studies and classic pedagogic theories, answers of curious questions were remembered better than those of boring ones (70% vs 54% correct, a 16% increase).

Curiosity effects on the hippocampus

Curiosity enhanced memory and activated brain regions involved in processing reward, but these regions are not the brain locations responsible for storing the type of information contained in the Trivia questions/answers (which correspond to a semantic memory type that endorses information about meanings and concepts). How, then, an activation of reward systems by curious questions potentiated semantic memory? To answer this, researchers analysed the activity in the hippocampus(g), a brain hub involved in several types of learning. Could curiosity modulate hippocampal activity?

Schematic of the experiment during the brain scan session. Participants see a question, which then disappears, and seconds later they see the answer. The latter disappears and then the cycle starts again with a different question. | Credit: Modified from Gruber (2014).
Figure 2. Schematic of the experiment during the brain scan session. Participants see a question, which then disappears, and seconds later they see the answer. The latter disappears and then the cycle starts again with a different question. | Credit: Modified from Gruber et al (2014).

Strikingly, hippocampal activity during the anticipatory phase (before the answer) was predictive of a successful memory for curious questions, although curiosity did not overtly modulate the activity in this structure during the subsequent stage of memory acquisition (while participants were seeing the answer). So, it looked like the hippocampus was “distracted” during boring questions so as its activity could randomly lead or not to successful memory. But during curious questions, hippocampal activity became aligned to the efficiency of subsequent memory acquisition, i.e., high/low activity lead to high/low probability of remembering the answer, respectively.

Curiosity or attention?

To discriminate whether the observed effects were due to real curiosity and not by mere enhanced attention, researchers introduced incidental information between questions and answers during the scanner session. They speculated that curiosity should enhance learning of any information being presented around that time, while attention (which narrows down the focus to a particular stimulus) should produce the opposite effect.

Full schematic of the experiment during the brain scan session. Participants see the question, which disappears. Seconds later they see an incidental information (a face), which again disappears. Later they see the Trivia answer. Then the cycle re-starts with different questions and faces. | Credit: Modified from Gruber et al (2014).
Full schematic of the experiment during the brain scan session. Participants see the question, which disappears. Seconds later they see an incidental information (a face), which again disappears. Later they see the Trivia answer. Then the cycle re-starts with different questions and faces. | Credit: Modified from Gruber et al (2014).

Thus, participants were exposed after the Trivia question to an image unrelated to the question, a neutral face, and memory for these faces was also assessed later during the post-scan memory test. Interestingly, participants also remembered better those faces presented after curious questions than those after boring questions. Yet small, the effect was significant (42% vs 38% correct) and persisted one day later (only for highly confident remembered faces, 35% vs 31% correct). Clearly, memory for incidental information was not impaired, suggesting a real curiosity rather than pure attentional processes.

While seeing these faces, curiosity produced subtle but significant changes leading to a correlation in the activity of SN/VTA, the hippocampus and improved memory. Also, functional connectivity between SN/VTAand hippocampus was increased.

The hippocampus-Accumbens-VTA loop subserving memory

These findings offer important evidence supporting a role in human learning for the Hippocampus-Accumbens-VTA loop 3. In this model, the hippocampus acts as a detector of relevant stimuli and then promotes a dopaminergic neuromodulation over itself to facilitate the storage of such relevant information. Studies in rodents and primates point that the hippocampus can trigger dopamine release by affecting the VTA (where source dopamine neurons are located) via a polysynaptic circuit passing through the Accumbens. Interestingly, dopamine on the hippocampus in rodents potentiates learning and induces enduring synaptic plasticity changes, such as Long Term Potentiation, which may contribute to the cellular storage of memory.

Diagram of the hippocampus-Accumbens-VTA loop (including other functionally interconnected regions) in the mouse brain. The hippocampus sends an excitatory (glutamate) projection to the Accumbens. The latter sends an inhibitory (GABA) projection to a region that in turns inhibits the VTA, thus releasing a blocker of VTA activity. As a result, the projection from the VTA releases dopamine in the hippocampus | Credit: Russo & Nestler (2013).
Diagram of the hippocampus-Accumbens-VTA loop (including other functionally interconnected regions) in the mouse brain. The hippocampus sends an excitatory (glutamate) projection to the Accumbens. The latter sends an inhibitory (GABA) projection to a region that in turns inhibits the VTA, thus releasing a blocker of VTA activity. As a result, the projection from the VTA releases dopamine in the hippocampus | Credit: Russo, S.J. and E.J. Nestler, The brain reward circuitry in mood disorders. Nat Rev Neurosci, 2013. 14(9): p. 609-25. DOI:10.1038/nrn3381

Researchers now proposed that human curiosity promotes learning by stimulating a dopaminergic neuromodulation of hippocampal activity. The activity patterns they found in the hippocampus, Accumbens and SN/VTA during states of curiosity vs non-curiosity and how they link to learning performance is rather consistent with this model.

Implications and limitations

Readers should be warned of two typical risks when interpreting brain imaging studies. One is the extreme reductionism. Although the study focuses on three brain regions in which activity changes in response to curious information, it does not follow that those regions are silent in absence of curiosity, or that other brain regions beyond these three are not activated during curiosity states. Researchers of this study never stated so.

In line with this concern, a critique that applies to most fMRi studies of cognition is: how relevant are these experiments to daily learning phenomena? We all routinely find curious and boring information, but very unfrequently these events consist of Trivia questions and even less they occur within the context of a brain scanner, laid down with restricted movement, etc. Researchers acknowledge this limitation. In fact they think their task under captures the spontaneous intrinsic motivation and curiosity of people. This could explain why they see such a small effect of curiosity on incidental learning of faces. But until we don´t have a technology allowing brain imaging in freely moving humans, we have to comply with fMRI studies, whatever is happening in there.

A second risk when interpreting many fMRI studies, including this one, is that they offer correlative evidence. Researchers showed here that activity in some brain regions is associated with states of curiosity and learning efficiency. But correlation never demonstrates a cause-effect relationship.

If these findings prove to reflect causal mechanisms, they would have several important implications. Clinically, a number of cognitive impairments (aging, Parkinson, etc.) are associated with dopaminergic dysfunction, suggesting that changes in the motivation to learn could underlie the memory deficits in them. In schools, these findings would support the view that stimulating curiosity beyond a mere acquisition of knowledge could significantly improve the academic performance of students.

References

  1. Weigmann, K., Educating the brain. The growing knowledge about how our brain works can inform educational programmes and approaches, in particular, for children with learning problems. EMBO Rep, 2013. 14(2): p. 136-9. DOI: 10.1038/embor.2012.213
  2. Gruber M. & Charan Ranganath (2014). States of Curiosity Modulate Hippocampus-Dependent Learning via the Dopaminergic Circuit, Neuron, 84 (2) 486-496. DOI: http://dx.doi.org/10.1016/j.neuron.2014.08.060
  3. Lisman, J., A.A. Grace, and E. Duzel, A neoHebbian framework for episodic memory; role of dopamine-dependent late LTP. Trends Neurosci, 2011. 34(10): p. 536-47. DOI: 10.1016/j.tins.2011.07.006

0 Comments

2 Trackbacks

Leave a reply

Your email is never shared. Required fields are marked.

Required
Required
Required

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>