The amygdala, an almond-shaped brain structure known as the seat of emotion, also contributes to higher thinking—including those “Aha” moments when we suddenly see the answer to a puzzle, a new study suggests. From functional magnetic resonance imaging (fMRI) analyses reported in the March 10 Neuron, Nava Rubin of New York University and colleagues found that amygdala activity during an instance of induced insight could predict, to some degree, participants’ long-term memory of the solution. Though ramifications for Alzheimer’s disease are unclear, the study suggests a larger role for the amygdala in memory than may have been previously appreciated.

“This article uses a clever design and nicely demonstrates the involvement of the amygdala in memory formation, especially during moments of surprise,” Trey Hedden of Massachusetts General Hospital, Boston, wrote in an e-mail to ARF.

Extensive research in rats and people has established that activation of the amygdala promotes formation of long-term memory, particularly of emotional experiences (see Hamann et al., 1999; for reviews, see Phelps and LeDoux, 2005; Cahill and McGaugh, 1996; McGaugh, 2004), but also less arousing ones (Roozendaal et al., 2008). For the current study, Rubin worked with first and second authors Rachel Ludmer and Yadin Dudai of Weizmann Institute of Science, Rehovot, Israel, to design experiments that would tease out brain mechanisms influencing memory of induced moments of insight.

The researchers showed participants 40 hard-to-recognize camouflaged images, eliciting after each an “Aha” moment by briefly presenting the original real-life picture (i.e., the “solution”). They showed the camouflage images again one week later, and gauged long-term memory by whether participants could correctly identify the embedded object without being shown the original picture. Functional MRI brain scans revealed several brain regions—most prominently, the amygdala—correlating with long-term memory of the solution. The researchers used the amygdala fMRI data to predict subsequent memory of the images in a separate group of participants. However, Hedden noted that the predictive power is “not especially strong,” correctly predicting memory outcome 62 percent of the time (50 percent is chance level).

Activity in visual, medial frontal, and parietal cortices also associated with memory retention. However, no such correlations showed up in the hippocampus or other memory areas of the medial temporal lobe that typically fade in AD. Lack of activity findings in neuroimaging studies are hard to interpret, though, Paul Reber of Northwestern University, Evanston, Illinois, pointed out in an e-mail to ARF. Furthermore, the authors do not report whether memory-related fMRI activity in the medial frontal and parietal areas “could also be used, either alone or in concert with the amygdala activity, to predict memory,” Hedden wrote. The data also leave open the possibility that the hippocampus, despite showing no memory-related fMRI effects, “could be involved in linking the camouflage image to the solution image that leads to later memory, and that this linkage is modulated by the observed amygdala activity,” Hedden noted (see full comment below).

Reber agreed that the results are less likely to suggest “a new role for the amygdala in memory formation,” but probably reflect known interactions between the amygdala and medial temporal lobe (MTL) areas. He is not sure the results easily apply to AD. However, given the relative sparing of the amygdala compared to hard-hit nearby memory areas, “we could stretch a bit to hypothesize that positive emotional experiences could help with memory formation in patients, perhaps especially sudden ones like insight,” he noted. Moreover, since the insight came from a cognitive task instead of other positive rewards such as money, food, or sex, the current data suggest “a wider range of ways to create the positive experiences that might help memory,” Reber wrote (see full comment below). “This effect might be more pronounced earlier in disease progression when there is less neuropathology in the traditional MTL memory system. That is, when the system is at least partly functional, it can be boosted by the insight response.”—Esther Landhuis

Comments

  1. While the amygdala has previously been associated with enhancement of long-term memory formation, this has been thought to reflect an emotional component—events that are emotionally important could be better remembered. The article by Nava Rubin and colleagues reports a similar effect based on inducing an "aha!" moment via a clever perceptual insight.

    In their report, the role of the amygdala is clear and appears to affect perceptual processing to enhance memory. No differential activity is found in the traditional medial temporal lobe (MTL) memory areas that are impacted in Alzheimer’s disease, but this is difficult to interpret, as lack of activity findings tend to be ambiguous in neuroimaging. Though it is possible the results reflect a new role for the amygdala in memory formation, more likely they reflect known amygdala-MTL interactions. It is still very interesting to see the amygdala engaged in an insight-driven paradigm where the emotional involvement is positive and based on a cognitive process.

    I'm not sure if the results are easily applicable to Alzheimer’s. Given that the amygdala is not as impacted in AD as the nearby memory structures, we could stretch a bit to hypothesize that positive emotional experiences, especially sudden ones such as insight, could help with memory formation in patients. Since the insight here comes from a cognitive (perceptual) task instead of other positive rewards (e.g., money, food, sex), it suggests a wider range of ways to create these positive experiences that might help memory. This effect might be more pronounced earlier in disease progression when there is less neuropathology in the traditional MTL memory system. That is, when the system is at least partly functional, it can be boosted by the insight response.

    View all comments by Paul Reber
  2. This study uses a clever design and nicely demonstrates the involvement of the amygdala in memory formation, especially during moments of surprise. Of particular interest is the possibility of predicting which images will be remembered one week later on the basis of amygdala activity during the "aha” moment of seeing the solution. However, it should be noted that this predictive power is not especially strong, making a correct prediction (actually remembered when predicted and actually forgotten when predicted) about 62 percent of the time (50 percent would be chance level). It is not clear whether activity in the amygdala alone would be the best predictor of later memory, as the authors don't report whether the memory-related activity in precuneus, medial prefrontal cortex, or anterior cingulate cortex could also be used (either alone or in concert with the amygdala activity) to predict memory.

    Although no memory-related effects were observed in the hippocampus during the time the solution was shown, there remains the possibility that the hippocampus could be involved in linking the camouflage image to the solution image that leads to later memory, and that this linkage is modulated by the observed amygdala activity.

    Although the amygdala does demonstrate atrophy during Alzheimer's disease (Cavedo et al., 2011), amygdala-dependent memory is not especially impaired during aging or dementia, although there are some indications that amygdalar modulation of hippocampal-dependent memory may be impaired. For example, the few studies (Hamann et al., 2002; Hoefer et al., 2008) that I know of examining fear conditioning (which relies heavily on amygdala involvement) in Alzheimer's dementia in humans use a particular form of conditioning known as a "partial reinforcement schedule," meaning that the person is required to learn that a neutral stimulus precedes an aversive event even though the neutral stimulus is sometimes presented without the aversive event (that is, the predictive pairing is less than 100 percent). Compared to a "full reinforcement schedule" (i.e., the predictive pairing = 100 percent), the partial reinforcement schedule likely requires more hippocampal involvement in learning the link between the neutral and aversive events. That is, even though these studies indicate a potential role for impaired amygdala function in memory failures during dementia, they do not rule out that this impairment acts through the amygdala's connections to the hippocampal memory circuits known to be impaired in Alzheimer's disease. It is also possible that the cortical regions (precuneus and medial prefrontal cortex) that show memory-related effects in this study would be impaired in Alzheimer's patients, preventing the ability to benefit from "aha” moments, as these cortical regions are the same ones that tend to accumulate amyloid plaques early in the disease process.

    References:

    . Local amygdala structural differences with 3T MRI in patients with Alzheimer disease. Neurology. 2011 Feb 22;76(8):727-33. PubMed.

    . Impaired fear conditioning in Alzheimer's disease. Neuropsychologia. 2002;40(8):1187-95. PubMed.

    . Fear conditioning in frontotemporal lobar degeneration and Alzheimer's disease. Brain. 2008 Jun;131(Pt 6):1646-57. PubMed.

    View all comments by Trey Hedden
  3. The research can become useful for therapeutic purposes or cognitive training. The pre-training involves an ambiguous image followed or paired with the actual original image. There is an emotion factor to consider. The ambiguous image can set up frustration when research participants may be trying hard to decipher “what it is.” There are also elements in the different places on the pictures that the participants would decipher as “where it is.” The integration of object things with their spatial location are two elements that would need to be integrated and matched, i.e., from the “memory trace” of the camouflaged figure to the original image. This mapping from an obscured image on to a clear image seems to involve stimulus generalization only, or matching to sample. Insight seems to require a third element. That involves use of disparate objects to reach another object, goal, or solution.

    The “what it is” process and “where it is” process are not mapped in a similar way when 1) matching (or alternatively a generalization of) an obscure sample image to a clear, related sample image, and 2) matching two disparate, unrelated objects. Two different, unrelated “whats” and “wheres” would have to be deciphered towards a goal (i.e., solution to a problem), or associated, for "it" to be considered an insight.

    Epstein and colleagues (Epstein et al., 1984) showed that insight can be as a consequence of stimulus and response training involving bringing two disparate objects together towards a goal or solution. The video is instructional to watch.

    References:

    . 'Insight' in the pigeon: antecedents and determinants of an intelligent performance. Nature. 1984 Mar 1-7;308(5954):61-2. PubMed.

    View all comments by Kiumars Lalezarzadeh

Make a Comment

To make a comment you must login or register.

References

Paper Citations

  1. . Amygdala activity related to enhanced memory for pleasant and aversive stimuli. Nat Neurosci. 1999 Mar;2(3):289-93. PubMed.
  2. . Contributions of the amygdala to emotion processing: from animal models to human behavior. Neuron. 2005 Oct 20;48(2):175-87. PubMed.
  3. . Modulation of memory storage. Curr Opin Neurobiol. 1996 Apr;6(2):237-42. PubMed.
  4. . The amygdala modulates the consolidation of memories of emotionally arousing experiences. Annu Rev Neurosci. 2004;27:1-28. PubMed.
  5. . Noradrenergic activation of the basolateral amygdala modulates consolidation of object recognition memory. Neurobiol Learn Mem. 2008 Oct;90(3):576-9. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Uncovering camouflage: amygdala activation predicts long-term memory of induced perceptual insight. Neuron. 2011 Mar 10;69(5):1002-14. PubMed.