Finding the self in the brain

We have just published an article in NeuroImage that explores ideas related to the self and how they are represented in the brain.[1. You can find the pdf at the journal’s site, or get it here] It was an enjoyable article to write: the self is such an inherently interesting subject. It is at once so familiar, but curiously enigmatic. The self — the concept of the self — becomes harder to grasp the more deeply you consider it. We didn’t delve too deeply into its epistemological aspects though, but rather into its brain aspects. We focused on how we could identify the core brain regions that support our concept of ourselves, and on how those regions interact in a network.

One of the major contributions that functional MRI has made to our understanding of the brain is in its identification of the default mode network. This is a network of regions that is more active when we are at rest — in our own thoughts — than when we are engaged in external cognitive tasks. It took some time to identify the DMN, because our interest was always in the brain activity related to the cognitive tasks. It was first recognised by an observant neuroscientist — Gordon Shulman — who, with his colleagues, saw that the same brain regions always showed reduced activity during the performance of tasks that required concentration and attention towards the outside world. And right from the beginning they considered that this network might support processes related to the self.

Not long afterwards, the group described the brain regions as forming the “default mode of brain function”: it is what the brain does when there aren’t more pressing events to attend to in the external or internal environment. And the default mode network was born; since their 2001 paper it has been discussed in over 3,000 articles.

Over 3,000 papers have been published on the DMN since the publication of “A Default Mode of Brain Function” (Raichle et al. 2001).

The brain regions that form the DMN largely overlap with brain regions that are active when we think about ourselves (papers discussing the self in terms of the DMN form the green patch in the figure). It shouldn’t be surprising that this is the case: when we are alone with our thoughts we spend much of that time considering past events, or anticipating events in the future, with ourselves as the central actor.

In our paper we set out to establish how thinking about the self was similar and different to being at rest. It was a simple task. Participants were presented with a list of personality adjectives and asked whether they thought the word described them (word such as “skeptical”, “perfectionistic”, and “lucky”), and at rest stared at a fixation cross (+). Our control condition used a matched set of words: participants were asked whether they contained four or more vowels. The figure below shows how thinking about the self and being at rest overlapped, and how they differed.

Patterns of brain activation during self-reference and rest show considerable overlap, but also distinct differences.

Perhaps most importantly, we wanted to identify the regions in the DMN which we considered to be the core regions for the self. We defined these as regions that were more active when participants were at rest compared to when they were engaged in a cognitive task; and further, that showed even greater activation when they thought about themselves compared to when they were at rest. We identified three regions that fit these criteria: in the medial prefrontal cortex (MPFC), posterior cingulate cortex (PCC), and left inferior parietal lobule (IPL). We then set about establishing how they operated in a network: how they influenced each other during self appraisal.

[caption id=”attachment_269" align=”alignleft” width=”300"]

We identified three regions that establish the structure of the self (A), and showed how they interacted in a network (B).[/caption]

We discovered that the sense of self originates in the PCC, which, with the IPL, creates a representation of the self that is projected forward to the MPFC. The MPFC shapes the representation of the self via its negative influence on the PCC; according to an appraisal of the external context. A metaphor: the PCC (with IPL) acts like the film producer, establishing the space for the shoot, where the MPFC acts like the director, focusing attention on those aspects of the space that are relevant to the broader picture.

So the self, in our estimation, is supported by a brain network that sits within the broader DMN. Our model has simplified it (as models do), but we think it provides some interesting insights into how the self is structured. At least at a brain level. And we are keen to explore how this might help explain what is going on in an illness like depression, where the sense of self becomes so disturbed.