Does the quality of “Smart Information Processing” connect with the Default Mode Network ?

I was triggered by an article in Nature, explaining the atypical connectome hierarchy in ASD (Autism Spectrum Disorder).

ASD is characterized by atypical sensory processing, while and deficits in high-level cognitive and social functions, including impairments in Theory of Mind and predictive abilities. The article points out that ASD might emerge from disturbances in macroscale cortical hierarchy. Notably, ASD display an initial processing of information from sensory, unimodal to early heteromodal systems (e.g. pMTG and OT). Transitions into transmodal regions (such as mPFC and PCC/PCU) were aberrant, and failed to converge on the DMN core.

This is an interesting observation, giving rise to some reflections I wanted to share.

Human evolution

Yuval Noah Harari described in his book Sapiens that humans dominate the planet not because we are logical, but because we had a cognitive revolution, which happened some 70 kyr or so years ago. During this CR, we developed the ability of creating stories, we agree to believe (and we know are untrue). These stories and beliefs are about organization (orders in the society, economic and political) and religions. Gossiping played an important role in this.

I want to link with Robert Sapolsky’s book, Behave (which he summarizes in this lecture). Apart from his focus on the importance of context, the dopamine story is of interest. Dopamine fuels the goal-directed behavior needed to gain a reward; it “binds” the value of a reward to the work. It’s about the motivation arising from those dopaminergic projections to the PFC that is needed to do the harder thing (i.e., to work). In other words, dopamine is not about the happiness of reward. It’s about the happiness of pursuit of reward that has a decent chance of occurring. The DMN core nodes, such as the mPFC, are among the most consistently activated regions during self-referential and introspective cognition as well as other-oriented cognitive perspective-taking operations including mentalizing.

He also tells ehere are no discernible differences between how the monkeys and people’s neurochemistry functions. The only distinction is that monkeys don’t get hooked on beliefs, ideologies, dogma, degrees, titles, fantasies, lies, empty promises, or self-deceptions.

Or as I once heard in a personal discussion with Frans De Waal:

“human man-kind is great in fooling himself, building illusions of what is not there”.

The link with anther DMN core node (PCU) is described by Daniel Goleman in Focus: The Hidden Driver of Excellence. Goleman persuasively argues that we must learn to sharpen focus if we are to survive in a complex world. He delves into the science of attention in all its varieties, presenting a long overdue discussion of this little-noticed and under-rated mental asset in an era of unstoppable distractions. This book helped me to discover the importance of the PCU and its relation to context and system blindness. In his book, Goleman explain these connections with PCU to become and essential requirement for leadership skills, like “inner focus” and “self-awareness”.

These authors are samples of references to the human identity as related to core nodes and functions of the transmodal DMN, in which we make higher abstractions and model of what is going on around us, build structures and stories, and project a future. And on top, we want to act in order to realize or evolve towards a realization of that future. A pretty unique feature humans have, compared to the other species of the world. 

When we project a future, we want to work towards “the good”. As stated by Yangmei Luo et al. , “what is a good life and how it can be achieved is one of the fundamental issues. When considering a good life, there is a division between hedonic (pleasure attainment) and eudaimonic well-being (meaning pursuing and self-realization).” Their paper states that the eudaimonic and hedonic balance (EHB) was positively correlated with functional connectivity of bilateral ventral medial prefrontal cortex (vmPFC) within anterior DMN and bilateral precuneus (PCU) within posterior DMN.

Towards the arts

At its the core essence, arts is ‘a story’. This story often contains a mixture of illusions, open ends, uncertainties, unknowns, multiple interpretations and so on. (Star Wars lovers, think of George Lucas as an artist – that’s how he describes himself).

In the world of music, this story is frequently being told with no words attached to it. Music activates your whole brain and practicing music changes your brain up to the level a scan can detect if you are a musician; some of the well-known facts how arts influences us. I want to reference Levetin’s book This is Your Brain on Music, and liked especially where he refers to his conversation with Nobel prize winner Francis Crick (DNA), who advices: “Look at the connections”. 

Nikolaus Harnoncourt once told in an interview “after the concert I feel I’m happy but not satisfied”. The lack of being satisfied creates the need to reach further and work harder. This links very well with the EHB balance. Most musicians do recognize this double feeling. 

In the artificial intelligence

Shimon Ullman recently stated in Science: “The superiority of human cognitive learning and understanding compared with existing deep network models may largely result from the much richer and complex innate structures incorporated in the human cognitive system.” Or, AI (and deep learning models) is just acting in a unimodal way with too little connections to other decision systems. This does not compare to what makes us human; the interconnection of this all into a transmodal network.

And even for the unimodal, we have to be careful, as mentioned in the article Unmasking Clever Hans predictors and assessing what machines really learn. Learning machines often behave very differently from humans. It is instrumental in verifying the correct behaviour to explain the decision making. Existing systems have a surprisingly rich spectrum of behaviours ranging from strategic decision making to ‘Clever Hans’ strategies or undesired behaviours. These different behaviours go unnoticed by common evaluation metrics, which puts a question mark to the current broad and sometimes rather unreflected usage of machine learning in all application domains in industry and in the sciences.

As this article is questioning the decision-making strategies of AI & ML, we still have no clue about the ‘core values’ the AI or deep learning has. Will an AI go for delayed gratification, or save the life of it’s child?

In conclusion

Humans live in an increasingly complex world. The raising level of information we have to process is putting us to under stress, or as Herbert Simon puts it:

“What information consumes is rather obvious: it consumes the attention of its recipients. Hence a wealth of information creates a poverty of attention, and a need to allocate that attention efficiently among the overabundance of information sources that might consume it.”

This urges us even more to “take time”, enough to activate the transmodal network, time to develop our attention and focus, time to develop valued good solutions, develop our DMN and make good use of it.

Sometime DMN is in my mind linked with “Do More Nothing” (i.e. refrain from immediate actions and activities, but take time reflect, sleep over it or even rest – and yes, the good functioning of the DMN is challenged by sleep deprivation).

It is well known, great thinkers and artist like(d) to walk or practice music, in order to refrain from thinking and overthinking. Many notes have been made on the creative results of these and similar activities of reflection.

Maybe – in this upcoming world of AI – we should not just walk on our own, but also ask our AI system to take a break and reflect on the good nature of its activities.

Maybe we should work towards AI systems we can ask to walk (our dog?) or hum some music, before deciding upon what is valuable to us.