Intelligence, Creativity and Information Processing in the Human and Synthetic Minds

What are the unique features of the primate brain that enable their higher intelligence?

Dr Alessio Delogu, Professor Sandrine Thuret, Professor Robert Hindges, and Dr Michael Berthaume.

The scientific question: could the unique abundance of inhibitory interneurons in the thalamus help explain primate intelligence?

Definitions:

• Intelligence is the cognitive ability that includes attention, working memory, and decision taking
• Primate is a group of mammals, including humans and monkeys, with large brain-to-body ratio. Primates are capable of high levels of intelligence.
• Thalamus sits at the center of the brain where it controls the flow of information required to generate thought and action. It exists across different species, but is most complex in primates. 
• Inhibitory interneurons are a subtype of brain cells that balances various signals going across the brain. They can be found all over the brain, including the thalamus. However, the primate thalamus contains 100 times more interneurons than other species, the only area of the brain with such striking difference.

The investigation:

Dr Delogu and Professor Thuret are working together to turn stem cells into these specialised cells in the lab. But first, they need to identify what factors (genetic or environment) are essential for stem cells to take this specific path to become inhibitory interneurons in thalamus. As these cells are only present in the thalamus, they need to be able to generate these cells in large numbers to study their unique properties.

Prof Hindges will investigate whether it is possible to generate a human-like thalamus in transgenic fish to study changes in their behaviour. Dr Berthaume will create a mathematical model to track the evolutionary history of inhibitory interneurons of the thalamus in species with varying levels of intelligence.

The expected outcome:

The primary goal is to develop a robust protocol to create these cells for all future studies. This protocol needs to be robust and repeatable.

One future study includes looking at whether altering the number of these cells affect cognitive ability. Comparing across different species, they will also be able to track the evolution of brain networks that contribute to the emergence of intelligence.

Pinpointing the source of intelligence in human can inform neuromorphic computing, a field where we apply insights from how brains process information into designing better and faster computers.

Computer engineers have not, to date, looked at the significance of inhibitory control for neuromorphic computing. Interdisciplinary expertise is crucial to address this question.

Can we teach AI to be funny using memes?

Professor Lorenzo Zucca, Professor Luca Viganò, Professor Gabriele Costa.

The scientific question: How does creativity, imagination, and uncertainty function in human and synthetic minds?

Definitions:

• Creativity is the ability to generate novel and valuable ideas or works through the exercise of imagination
• Imagination is the ability of forming a mental image of something not present to the senses or never before wholly perceived in reality. It is also a way to deal with uncertainty
• Uncertainty is a situation where there are imperfect or unknown information.
• Memes are internet content that combine images and text. Memes are unique as they represent creativity and imagination, and are products of rapid cultural evolution. They are useful in this study as humour (and comedy) rely on human spontaneous creativity. Memes are created by human, but their distributions are curated by algorithm tailored to user preferences. The Kaggle meme dataset has gathered over 7,000 internet memes. It is used for training machine learning models in computer vision (image recognition) and natural language processing (text analysis)

The investigation:

Scientists will systematically analyse memes to develop metrics of creativity.

They will also design neuroimaging protocols to study brain activation when interacting with memes. Furthermore, they will examine memes’ virality to study the role of algorithm in recognising and distributing creative content.

The expected outcome:

The main goal is to produce metrics of creativity through a philosophical, neurological and computational lenses. 

This will provide models for AI creativity, where scientists can train AI to better understand the criteria of creativity as humans understand them. Scientists will also gain insight into the potential and limitation of AI creativity.

Furthermore, scientists want to understand the misuse of creative expression and how creativity may be used for propaganda communications.

Is it possible to record the activity of an entire brain?

Professor Robert Hindges, Professor James Millen, and Dr Flavio Dell'Acqua

The scientific question: Can we capture whole-brain neural activity using a neuromorphic camera?

Definitions:

• Neural activity describes how electrical signals travel through the brain circuits, allowing the brain to process information.
• Neuromorphic camera is a bio-inspired camera mimicking the working of human eye. A neuromorphic camera only records data when brightness changes. They are also referred to as event-based cameras as they only turn on when they detect an “event”. They have been used for motion detection, robotics, and scientific imaging.

The investigation:

Understanding how brain process information requires observing the precise timing of electrical signals across the entire brain network.

Traditional imaging technologies are currently not sophisticated enough to capture every electrical event in the brain. They have to choose between capturing 100% events of one single neuron, or capturing only 2% events when looking at thousands of neurons.

Scientists want to improve neuromorphic cameras to be able to detect all events across the whole brain. This has never been done before.

The expected outcome:

The main goal is to produce a highly-sensitive and ultra-fast imaging technique that is able to capture all electrical events across the whole brain in real time

Professor Hindges has successfully developed a novel zebrafish line that allows for whole-brain electrical signal imaging. Professor Millen has developed neuromorphic imaging to track microparticles. Dr. Dell’Acqua will develop the methodologies to extract and analyse the neuronal signals from the data.

Combining their interdisciplinary expertise, the team will create a novel way to capture neural activity across the entire brain of zebrafish. This will help dissect signal processing in the human brain and better understand its complex function.