On the third night I woke to find my body filled with a strange energy. It felt like soda water fizzing beneath my skin. At first I was excited. I sat up in my bed and wondered what would happen next. But nothing happened next. The energy didn’t go away or change. It was just there. After about twenty minutes I was bored. I wondered how long it would last. It was midnight and in five hours’ time someone would bang a gong outside my door. Then I would rise, drink tea in the communal kitchen with the other guests – in silence; this was a silent retreat – and then we’d march uphill through the rain and dripping trees to the meditation hall, a classroom-sized room with bare wooden floors, white plaster walls, and windows looking out over a deep valley. We’d meditate for two hours, then break for breakfast, then sit for another two and a half hours, then two more hours in the afternoon and one in the evening. I knew that if I didn’t sleep I’d spend most of the next day dozing or daydreaming instead of meditating. A whole day of my five-day retreat wasted. I lay on my back, seething. When, I wondered, steepling my fingers and drumming them against my chest, would this mystical experience end? I tried to relax. I focused on my breathing. Eventually I fell asleep.

I woke feeling normal. I got up, drank my tea and formed one of the vague shadows moving through the pre-dawn dark up the trail to the hall. Someone had already lit the fire. The rain was steady on the roof. We sat cross-legged on mats or wooden meditation stools or chairs, with blankets draped around our shoulders. Our teacher chanted in Tibetan while the rest of us mumbled along. Then we meditated. About forty minutes in, I was alert but calm – no thoughts, no distractions – when I perceived an object in my mind, hovering at the periphery of my awareness. It was a greenish-gold colour, like iridescent scales. It had a high musical tone, a vaguely circular shape. I switched my attention to it and it amplified, slowly at first then quickly: a rogue algorithm, looping, asymptotic, exploding inside my mind and pulsing off and on, a strobing effect blotting out all self-awareness, all other sensations. It was both pleasurable and frightening. I had no conception of my body yet at the same time I felt like I was falling, faster and faster. I couldn’t breathe.

*

What is enlightenment? It’s still not clear to me. In traditional Buddhism, it’s linked to the pre-Buddhist Hindu belief in karma and reincarnation. Enlightenment is how you escape the endless cosmic cycle of suffering and rebirth. But most contemporary Buddhists seem to prefer the term ‘awakening’, the metaphor being that we’re all asleep and don’t know it, and this confused state is – allegedly – the cause of our suffering. When we wake to the true nature of reality, the suffering ceases, revealing a nameless, unimaginable form of freedom – a new perception of existence, some state beyond words.

There’s a lot of – mostly – amicable disagreement about what awakening is like and whether anyone has ever really achieved it. Some argue that there are multiple stages of awakening, like levels in a video game; others say that awakening is just an ideal, something to aspire to in your meditation practice but which you can never really reach.

There’s more consensus around the jhanas, the alternate states of consciousness that precede and point towards awakening. The story goes like this: when the Buddha was a young man he left home to become a mystic. He studied under several spiritual masters who taught him yoga and meditation and these skills allowed him to access the jhanas. If you reach a deep enough level of concentration, you can access these progressively sublime and indescribable mental states, and the way you think and experience the world literally changes.

There are four jhanas, or maybe eight depending on who you ask. You enter them sequentially, and each requires a higher level of meditative skill to access. The Buddha reached the highest jhana, but was still dissatisfied, still not awakened, so he sat beneath a tree (the Bodhi tree, now a famous pilgrimage site) and vowed not to move until he achieved awakening. He did this, the scriptures claim, by going beyond the jhanas and perceiving the three characteristics of existence, deep insights into the true nature of reality.

Awakening is mysterious, but the jhanas are a real thing. Lots of people experience them and have done for thousands of years. The Digha Nikaya (‘The Collection of Long Discourses’), one of the scriptures in the Pāli Canon, the oldest extant Buddhist texts, describes the first jhana: ‘Quite secluded from sense pleasures, secluded from unwholesome states of mind, he enters and dwells in the first jhana, which is accompanied by applied thought and sustained thought with rapture and happiness born of seclusion. One drenches, steeps, saturates and suffuses one’s body with this rapture and happiness.’

This sounded good to me, and about a year into my meditation practice I started to follow the instructions to jhana access. But being secluded from sense pleasures and unwholesome states of mind is hard work. Or, rather, it is hard work to get to the point where it no longer feels like hard work, which is the point you need to reach.

You need sustained and effortless access concentration in order to enter the jhanas, and here is where we meet the famously paradoxical nature of meditation: if you want to achieve a jhana state, you can’t. Only by not wanting it can you obtain it. The way this worked for me is that I meditated for hours until I finally found myself in a calm, thoughtless state, effortlessly focused on the breath, and then my mind blared, ‘I’m doing it! I’m so calm and focused!’, shattering my equanimity and flooding my thoughts with feelings of disappointment, frustration and exasperation, which are of course additionally destabilising.

I persevered. But I still couldn’t enter the jhanas. I felt blocked. I had weird pulses of energy in my hands during the day. At night I experienced a fluttering in my mind, like something trying to batter its way free. I wasn’t relaxed enough, the texts suggested. My main source for studying jhana practice – Leigh Brasington’s Right Concentration: A Practical Guide to the Jhanas – explained that it is hard for people living in the day-to-day world to access these states. You just don’t get the seclusion and calm you need for the enterprise. The Buddha himself provided advice on where to go for jhana practise. Avoid large or new monasteries, he counselled. Frontiers, ports and haunted places are also bad. Instead he recommended secluded dwellings, caves, clefts in cliffs. Cemeteries. And mountains.

*

The Wangapeka Study and Meditation Retreat Centre sits on the higher slopes of a large hill, arguably a small mountain, in the northern foothills of the Southern Alps. You reach it via a long and winding dirt road leading to a steep and winding dirt driveway, leading to a carpark and then a series of dirt trails through the trees. There are huts and campsites scattered about the clearings. Some of them have ‘Retreat in Progress’ signs posted on barriers blocking entry to their trails. These are occupied by solo retreatants: people who’ve been up there for months, meditating alone.

The main building, midway up the hillside, houses the kitchen, toilets and showers, and a large communal space that functions as a combination of dining room, meeting room and library. French doors open out onto a rectangular lawn where you can practise yoga or tai chi, if you are so inclined, which almost everyone who would stay at a place like Wangapeka is. At right angles to the main building is a row of narrow, cell- like guest rooms, the third of which was mine.

The centre was founded in 1975, and was built up gradually over the years by members of the community – mostly New Zealand Buddhists who followed the Tibetan tradition and lived in the region, of which there were surprisingly many – hauling wooden beams and bags of concrete mix up the slopes, clearing trees, hacking pathways through the bush.

There was a map of the hill posted in the foyer of the main building. It claimed to show all of these pathways, including routes to a lookout, a waterfall, a pagoda and a swimming hole. The retreat schedule allowed the guests a few hours of free time each day and I spent most of mine attempting to find these locations, but the map was decades out of date, so I always got lost. I wandered between the pines and tangles of wild blackberry, the long wet grass of the trails soaking my shoes, socks and trackpants up to my knees. Sometimes I encountered other equally lost wanderers, also on retreat. None of us were allowed to talk so we always shrugged theatrically and mimed a stream or waterfall: whatever we’d found, and the direction we’d found it. Pagodas are hard to mime.

On the third day, the day before I reached the first jhana, I found the lookout: a clearing on the north side of the hill. Far below it were farmlands – asymmetric light and dark green fields, the Wangapeka river winding between them, disappearing into a distant haze of sunlit mist. The fields were flanked by pine-covered hills, steep slopes of forest broken by cliffs, waterfalls spilling into the air, low clouds interrupted by columns of light.

I was in a philosophical mood. Partly it was all that silence and alpine air, but also it was having spent the months before the retreat trying to read Being and Time, a dense, often incomprehensible book by Martin Heidegger, who wrote extensively about anxiety and depression.

All through his life Heidegger tried to understand the problem of Being. Everything I saw from the lookout – the hills, the river, the light – existed in the world; they were all entities instantiated in space-time, and although they were all different, Heidegger pointed out that they all shared the same mysterious property of existence. What was this property? What was Being? How is it that we are part of the world, made up of the same matter and energy as everything else, but also able to observe it?

He felt that these questions lay outside the boundaries of rationality. They could never be explained by studying, say, physics or psychology. The sciences can explain the world of facts, he conceded, and there is nothing supernatural within the world they cannot describe. But sometimes, in moments of deep thought, or depressed or anxious moods, the world of facts and language seems like a thin membrane stretched over the unknowable true nature of things. We barely know what we really are and we barely know that we do not know. His philosophy depicts the world as a dark and endless forest in which humans are clearings in space and time, where existence suddenly opens up, glimpsing itself, and Being briefly emerges into the light.

*

Mood obviously has something to do with neurotransmitters: tiny chemical messengers that interact with neurons, the hundred-billion-odd branch-like elongated cells that make up our nervous system. The most famous neurotransmitter is serotonin, a fairly small, simple molecule. If you tell your doctor you have depression or anxiety they’ll probably prescribe an SSRI: a Selective Serotonin Reuptake Inhibitor, a class of drug that increases the extracellular levels of serotonin in key regions of your brain, and which sometimes improves mood and sometimes does not, for reasons nobody yet understands. The scientific debate on SSRIs is very heated. Some studies find they’re an effective treatment for depression and anxiety, others that they’re almost indistinguishable from placebos. Back in my late thirties, when my doctor first diagnosed me and wrote out a prescription, I’d already spent hours reading the clinical literature on antidepressants and obsessing over anti-pharmacology conspiracy websites. I explained all of my doubts about the drugs in between bouts of exhausted weeping. ‘Try them anyway,’ my doctor advised. ‘All I know is that they definitely help some people.’ So I tried them and they helped me.

We know the chemical structure of serotonin and the rest of the monoamine class of neurotransmitters that are linked to mood disorders. We know how they’re metabolised and transported around the brain. We know how they interact with our neurons. What no one understands is how all of this biochemical activity affects the experience of subjective consciousness. Why does serotonin make you happy? How does it affect mood? What is mood? What is depression? How does any of this stuff work?

Philosophers refer to these questions and the debate around them as ‘the hard problem of consciousness’. If you’re standing in a room and the room gets colder, a thermometer will measure the drop in temperature as the liquid inside it contracts as it cools. But humans and other species feel the drop in temperature. We have an internal, subjective experience of being cold. Our brains create this feeling, along with other feelings of pain, happiness, sadness, joy, etc., which philosophers refer to as qualia. But no one knows how the brain does this, or what the neurons are doing that is so profoundly different from thermometers and all the other matter in the universe. To put it another way: how could we reverse engineer consciousness and build a thermometer that feels the cold? Where would you start? Consciousness seems to be a property of the universe that everyone is familiar with – it is, if you want to get deep, the only thing we’re ever familiar with – but which is hard to even think about, let alone investigate scientifically.

Some philosophers say that the hard problem can never be solved (Heidegger would say it can never be rationally solved) or that the failure of biologists to solve it means that evolutionary theory is incomplete, possibly wrong. Others ‘solve’ it by declaring that consciousness is an illusion and that therefore there is no problem to solve. The psychologist Susan Blackmore points out, ‘An illusion is not something that does not exist [. . .] Rather, it is something that it is not what it appears to be.’ So even if there is an illusion, there is still a phenomenon to be explained.

Others insist that the question itself is incoherent and that by stating this they have solved the mystery. (Ever since Wittgenstein, it has been fashionable for philosophers to announce they’ve solved a problem by restating it in a formal language that proves it doesn’t exist.) It’s all a very lively debate. But from my perspective, depression feels like something is wrong with my consciousness. The way I experience the world is broken. Even if this is all just an illusion, it would be nice to know how the trick works, and why it fails.

*

There were about twenty fellow guests on my retreat and everyone had a job to do. Some people helped prepare the meals. Others cleaned. I washed the dishes after dinner. I resented this when I read about it in the booking confirmation email. ‘I’m paying these people money to wash dishes?’ But when you see the system in action it makes sense. A group meditation retreat is a very weird social situation: you’re living in close proximity with a bunch of strangers, but you’re not allowed to talk to them. The job system means social bonding still happens because everyone does things for each other, little reciprocal chores. So you feel like you’re part of a community, but there’s no gossip or status anxiety because there’s no conversation.

When I told my friends about the retreat they fixated on the commitment to five days of silence, wondering how anyone could stand it. I looked forward to it. I’m an introverted person, and this manifests as a sense of unease in most social situations. The unease is usually faint, a buzzing at the edge of my mind, easy to ignore, but if there are too many people around me the unease builds. Then I feel uneasy about feeling uneasy. Crowded pubs are stressful, as are buses and trains. Concerts. Airports. My daughter’s ballet school when all the classes change and I am surrounded by dense flocks of small girls dressed in pink, and their parents and bored siblings,  all yelling at each other while my family yells at me. In these situations the buzzing amplifies, drowning out everything else, and all I want to do is get away.

This introversion tips me towards unsociability. I don’t eat lunch with my workmates. If I have the choice between going out at night and staying in, I stay in. And even when I’m at home with my family, my inclination is to go into another room and read a book or look at the internet. At the same time I have a deep compulsion towards distraction, a dread of boredom. My mind always wants to be occupied, by a book or social media or a video game. Anything. There’s an infamous psychological study from 2014 in which people were given electric shocks then asked if they were willing to pay money not to be shocked a second time. If they offered to pay because they really hated being shocked, they were left alone in the room with the electroshock equipment still wired up, and over half of them – predominantly men – deliberately shocked themselves out of boredom.

I am very much part of this shock-yourself-for-no-reason demographic. Whenever I sit and meditate, a part of my mind screams at me that I have to open my eyes, have to go and do something, anything other than just sit and be quiet. I associate my depression with these two traits: introversion and distraction. I want to be alone but then I feel lonely, and I want to be distracted but then my brain feels overwhelmed. So I found it interesting that the mindfulness retreat seemed designed to prohibit exactly this pattern of behaviour. Distraction was banned and the environment was communal but minimally social. You weren’t lonely, but you were stuck with yourself and your thoughts.

When we first arrived at the retreat we were allowed to talk to each other, and the question most people asked me was ‘How long have you followed Jakob?’ Jakob was the teacher. He ran the retreat, and people were surprised to learn that I didn’t follow him, or even know who he was. ‘He was a student of Dilgo Khyentse Rinpoche,’ they explained in hushed tones, naming one of the most revered figures in Tibetan Buddhism, a teacher of the current Dalai Lama, but I didn’t know who Dilgo Khyentse Rinpoche was either.

Jakob was a lean ponytailed Norwegian in, I guessed, his late fifties or early sixties, based on his comments about backpacking around India as a teenager in the early seventies. He looked younger; his meditation posture was an effortless lotus. He gave us hour-long dharma talks every night. We could approach him whenever we had a question about his teachings, and each of us had a brief, scheduled meeting with him to discuss our spiritual progress.

My meeting took place in the Octagon. This was a small building just down the hill from the main lawn. It consisted of a single empty room with nothing on the walls, and windows looking out into a bank of white fog. Jakob waited inside, sitting on a chair in the centre of the room. He asked me about his talks. Did I understand everything?

This was awkward. I’d booked the retreat without knowing anything about the different strands of Buddhism, assuming any teacher could answer all my questions about jhanas, but on the first night Jakob had explained that jhanas weren’t part of the Tibetan tradition he followed. ‘It’s not my job to dangle alternate states of awareness in front of you,’ he’d said, before going on to tell us about lamas and reincarnation and higher realms of existence.

But I’d come on this retreat to try and access the jhanas, so during the meditation sessions I followed my own practice and ignored all of Jakob’s advice. This seemed like a rude thing to say to his face, so instead I asked him about consciousness and the hard problem, subjects that had been much on my mind. This was partly Martin Heidegger’s fault, but on retreat you do nothing but consider the self, and you inevitably wonder what you are observing, and what you’re observing it with.

‘Buddhism believes that consciousness is primary,’ Jakob replied. ‘Consciousness creates the material world.’

‘But the brain obviously creates consciousness,’ I replied. ‘How could consciousness precede the material world when it’s created by it?’

‘Is the conscious mind created by the brain?’ Jakob asked, smiling the smile of someone who is both filled with wisdom and loving compassion, and aware that he is about to kick someone’s ass in a spiritual debate. ‘How does it do that?’

I didn’t have an answer to this question. Nobody does. But  I didn’t worry about it for the rest of the retreat, because I didn’t worry about anything. I meditated. I stared at trees and clouds. I woke at night pulsing with energy. I felt astonished by the presence of Being. On the fourth morning I accessed the first jhana, felt my skull flood with light, fell out about twenty seconds later, and spent the rest of the day stabilising the state, resting in an uncomplicated mode of pure bliss. I spent my spare time walking the trails, feeling the bliss echo and reverberate inside my deliciously empty mind. I stopped at the sealed-off paths leading to the self-retreat cabins and wondered what it felt like inside the heads of the people who’d been up there for months.

But I thought about Jakob’s question on the flight home and I thought about it some more back at work where, after a day or two, all of the effects of the retreat evaporated and I felt identical to how I’d felt before I left. I meditated for hours in the early mornings and tried to access the jhana again but failed, gave up, and went back to my rudimentary half-an- hour-a-day practice.

I read essays and papers on cognitive psychology and neuroscience. The field was entering a golden age – staggering technological breakthroughs pointing to a grand unified theory of human cognition. This theory went by several names: predictive coding, or predictive processing, or the free energy principle. It was complex, highly mathematical and way beyond my comprehension. But it pointed towards the things I was interested in, like explanations of mood, mood disorders, tentative solutions to the hard problem, and an answer to Jakob’s question. When the summer rolled around and my university closed down for Christmas, I printed out a pile of neuroscience papers and took them on holiday with me.

We rented a place by the sea in Waikanae, the spiritual and geographic opposite of Wangapeka: a low-lying labyrinth of malls, golf courses, luxury car dealerships and retirement villages, all repeating with such regularity it seemed randomly generated, as if you could go on driving past the beaches and expressway on-ramps forever. Instead of meditating in the early mornings I walked along the beach, following the maze of paths through the sand dunes. Sometimes I wrapped myself in a blanket and sat outside on a recliner to read my predictive processing papers while the sun rose. Other mornings I just sat and looked at the sea. I started to feel both anxious and weirdly anticipatory about climate change. I fantasised about storm surges obliterating the baches and mansions arranged along the curve of the coast. I stopped meditating altogether. What was the point? And when the holiday was over and I was back at work, I found myself waking in the middle of the night, every night, and my depression settled in again.

Once when my daughter was a baby I hid her favourite toy behind my back. I didn’t know that you played this game with older babies but not young ones, because young babies don’t have object permanence. They assume that anything vanishing from their visual field no longer exists. As soon as the toy – I think it was a pig – was out of sight she started screaming.

A newly formed baby’s brain is a complex biochemical- electrical goo encased inside a small but expanding skull. It doesn’t understand much about the world, but there’s a constant stream of sense data – light, movement, smell, sound, touch, taste. It is far more information than can possibly be processed. The brain’s job is to take that torrent of data and build an accurate model of reality, so that as the baby grows it can navigate the world, move its body around in it, evade predators, find food and mates, survive amidst the deluge of information.

The way we do that, according to the predictive processing model of cognition, is by generating predictions and matching them against our sense data. If someone sticks their finger in a baby’s face, its brain will wonder: will this object feel solid if I touch it, or will my hand simply pass through it? The brain makes a prediction about how something will feel, look, sound or behave, then it tests the prediction against the incoming signal. If the prediction is validated, the assumptions behind it are built into the brain’s model (‘visible things feel solid’) and used to generate more predictions: ‘Given that this person’s finger is solid, will it taste good if I put it in my mouth and bite it really hard?’ This is why children like the same stories and games and experiences repeated over and over. If you’re building a model of reality based on prediction, you want the same datasets to fine-tune against. Humans have a much longer period of childhood than any other species, suggesting our predictive models are far more complex.

A few weeks after we got back from the beach holiday I was at home alone, feeling sleep-deprived and depressed. I walked from the unlit hallway into my bedroom and saw a very tall woman in old-fashioned clothes standing with her back to me, backlit by the window. It was very frightening. It also wasn’t real, just a suit jacket hanging from the curtain rail with the afternoon sun behind it.

Our brains never stop making predictions (‘the bedroom will be empty’) and matching them against the information our senses obtain about the world. If the prediction fails (‘There’s an impossibly tall woman in my room!’), it generates an error and places the brain in a state of statistical uncertainty. The brain hates this. Predictive processing argues that all human behaviour can be framed in terms of reducing error and uncertainty. If you walk into a dark room, it’s really hard for your brain to predict what’s inside it, so your immediate impulse is to turn on the light, because that makes it easier to check your predictions. And we don’t notice that our brain is constantly revising its predictions, because it conceals this process from our conscious awareness. A few days ago I was driving through the countryside and saw something in my peripheral vision: a dark object moving through a field. I turned to look at it, and saw, for a split second, a black lamb, then a black horse, and then the object stabilised as a human in a black jacket, along with the overwhelming impression that it had always been a human in a black jacket, that I hadn’t seen anything else. (Having the clarity to catch these little revisions as they happen is one of the dubious benefits of concentration meditation.)

Once we have a robust model of reality that seems to match the the world around us – meaning, it generates predictions with minimal error – our brains can filter out all the the irrelevant data the the world saturates it with. We have a model of language processing, for example, and one of the the predictions it makes is that the the multiple repeats of the word ‘the’ in this paragraph are errors, so it strips them out before presenting them to conscious awareness. And we often don’t notice the the sound of the air conditioning in an office, or the the traffic on distant roads, or other low-information noise because, again, our model of reality assumes it has no value. We might hear it now, or see redundant words if my mentioning these things means that the the prior assumptions have been temporarily challenged, but after a few minutes they disappear again.

The result of all this computation and filtering is the phenomenal world: the sequence of predictions that plays out inside our minds. When we wave our hands in front of our face, the image is happening inside our skulls. Our brains have no direct access to reality but, starting from nothing, they’ve constructed a detailed, controlled hallucination, calibrating it against changes in air pressure translated into sound, streams of electromagnetic radiation into images, sense data into touch. Even when we take action in the world, the brain makes a prediction – ‘I will scratch my nose’ – and the body moves the arm, hand and fingers to make the prediction come true. All we ever encounter is prediction upon prediction, the exquisitely calibrated hallucination; and all we ever feel, all our joy and pain and sadness, is a fluctuating measure of uncertainty about the statistical accuracy of the model.

It always starts with the insomnia. This isn’t so bad, at first. My bed is comfortable, and I never feel like I have enough time in the day just to think, so when I wake at 1am with inane thoughts and snatches of a song looping around my head I try to stay calm. If it is summer and the curtains are open I can see the tops of trees and the night sky and sometimes the stars. I pass the time by writing things in my head, or making plans for the future, or litigating arguments with my enemies on the internet. I take great care not to disturb my wife, and so long as I stay absolutely still and breathe regularly I am fairly safe. I feel awful the next day and always assume I’ll sleep well that night, but I wake up at 1am again. It only takes a few days of this before I realise I’m no longer tired; I’m depressed.

Depression is not always depressing. Sadness is not the default state. I mean, yes, sometimes you’re sad, but the classic symptom of clinical depression (and remember, no one knows what these diseases are; all we have are symptoms) is anhedonia: the inability to experience pleasure. Plath famously compared it to being trapped under a bell jar, walled off from the rest of the world. You feel nothing, and it sounds paradoxical but feeling nothing feels horrible.

I’m now familiar with the behavioural markers of anxiety and depression, at least in myself. The insomnia. The obsessional looping: the same thoughts, usually negative, self-critical, over and over again. I become withdrawn and antisocial, even more than normal. I want to stay inside with the lights off and the curtains drawn and binge-watch TV. I become irritable and selfish, even more than normal. I’m indecisive. I cry over trivial things, and during the post-beach holiday depressive episode I found myself blinking back tears one day when I tried to cross a busy road and there were no gaps in the traffic. I attempted to meditate but even though I could lie in bed watching sitcoms for hours I couldn’t concentrate on my breath for more than thirty seconds. I went back on antidepressants. This always feels like a miserable defeat and it always works. I felt better. I started meditating again and slowly tapered off the drugs. This took about two months.

*

Karl Friston, the world’s most-cited living neuroscientist, is the researcher most closely associated with predictive processing. When Friston was eight years old he spent a summer’s afternoon watching insects scurry around in his back garden in suburban England, trying to understand the purpose behind their activity. At first he thought they were seeking shelter and darkness because they spent more time in the shade, but then he realised this was selection bias. The insects in the sun moved faster than those in the shade because they were energised by the warmth. There was no purpose. The insect’s behaviour was just random.

Scientists prefer explanations that are parsimonious. They try to see the deep simplicity behind the apparent complexity of the world. Friston and other champions of predictive processing are excited by their theory because it explains so many diverse features of the brain that otherwise seem mysterious. What does consciousness do? What does the imagination do? Why do we sleep? What are dreams? What is mood and why do so many people have mood disorders?

The question of what consciousness does is not the ‘hard problem’, which demands to know how consciousness works. Instead it addresses a related debate about why consciousness even exists. What purpose does it serve for us to be able to ‘feel’ the cold instead of just registering it in some physical state, like a thermometer? To talk about this problem, philosophers invented a thought experiment: the famous philosophical zombie, or p-zombie.

The p-zombie does not look like a horror movie zombie. It looks and behaves exactly like a normal human. The only difference is that it has no consciousness. It experiences the cold in the exact same manner as a thermometer. A lot of human cognition takes place out of reach of our conscious awareness. While you’re reading this your brain is regulating your heart rate, blood sugar, oxygen levels, and countless other physiological states, and it’s making corrections, occasionally sending signals to your conscious mind telling you to eat something, or urinate, or go to sleep. You have no executive access to any of that information. You can’t shut your eyes and estimate your blood gas levels, but if your brain stopped measuring and regulating them you’d die very quickly.

All of this activity, Friston claims, consists of the brain making predictions about the body’s internal states and then minimising error between the prediction and the data. A p-zombie’s brain works the same way, only there’s no subjective awareness at all.

Two figures meet in a clearing. One is a human. The other is a p-zombie. They both ‘see’ a snake in the grass, but the human sees the snake in the same way all of us visualise the external world inside our own heads. The human feels afraid, screams and runs away. But the p-zombie doesn’t see or feel anything. Instead the image of the snake is projected onto its retina. The visual information is processed by the brain in the same unconscious autonomic way our brains process information about blood gas. It identifies the snake, matching it to images stored in memory associated with information about danger and threat, instructs the mouth and lungs to scream to alert anyone nearby, and issues more instructions to its skeletal and muscular system to turn and run away. There is never any fear, or any feeling or thought at all, only a flow of chemistry and light and information.

Back in the nineteenth century, the biologist Thomas Huxley took great delight in declaring that consciousness was an ‘epiphenomenon’: a pointless by-product of cognition, like exhaust from a car. Our awareness of our lives was, he claimed, ‘like the bell of a clock that has no role in keeping the time’. This was an amusingly shocking thing for Huxley to announce at public lectures in the late Victorian age, to audiences who felt consciousness was bound up with human exceptionalism and our immortal souls. But he had no proof for his claim and it seems not to be true. The consciousness of living creatures exists in layers of sophistication and complexity, the simplest of which are evolutionarily ancient, which suggests that consciousness is doing something vital in terms of natural selection.

The Portuguese-American neuroscientist Antonio Damasio identifies the seat of subjective consciousness in the brain stem, a primordial component of the nervous system. Maybe this is absurd, but I’m troubled by the idea of consciousness emerging long before the development of communication; that most of life on Earth consisted of aeons of sentient creatures awakening in the warm oceans, thrown into existence with no way to connect, living and dying alone, repeated trillions of times. This is the sort of thing I think about when I wake at 1am.

*

I grew up in a valley in the suburbs. Lots of families there had dogs, including ours, and sometimes a dog would start barking and all the other dogs would bark back, back and forth across the valley, then people would yell at the dogs, and more dogs would bark at them. This could go on for hours.

The Mind Illuminated – one of the most popular secular Buddhist texts – presents a model of consciousness drawn from the Yogacara school, a Buddhist philosophical tradition dating from the second to the fourth century CE. Like all good philosophers, the Yogacarins were interested in the nature of perception and cognition, and they argued that the reality presented to us in the mind is just a representation of the real (which aligns to the predictive processing claim that it’s all just a calibrated hallucination). But they also argued that the mind is modular – that instead of a single, unified self, it consists of many separate components, or subminds.

This is something that becomes apparent when I meditate. I sit and close my eyes because part of me wants to, but then another part of my mind wants me to get up and go do something else. I try to concentrate on the breath. My mind wanders. I bring it back. It wanders again. This is boring but, the Buddhists claim, it’s teaching us something profound: that there is no single, unified self. Consciousness is merely a way for the separate subminds to communicate with each other. All of our decisions are made at a subconscious level and then broadcast to consciousness. Sometimes the subminds disagree and we feel indecisive. Sometimes they clash, violently, and we feel torn. But our consciousness is only ever a witness, a screen for subminds to project thoughts onto. It has no agency. There’s no free will, no soul, no indivisible self.

Restated in the language of modern neuroscience, the brain consists of linked, large-scale networks, and consciousness is a global workspace for them to share predictions and responses. So one component – probably the fronto-parietal network, roughly the upper back part of the brain, which is associated with goal-driven behaviour and executive planning – decides to meditate and tries to focus on the breath. It frames this as a prediction and broadcasts it so the rest of the brain knows which movements to make to minimise the uncertainty and make the prediction come true. But when I sit and kneel and close my eyes, another component, probably the default mode network, a module involved in thinking about the past and predicting the future, activates and broadcasts its own inane, pointless thoughts, over and over, and then other networks – memories, emotions – respond to them. A non-mystical way to describe the early, frustrating period of learning to meditate is that you’re training all of these networks to shut up and stop barking at each other.

So the fourth-century Buddhist reply to the twenty-first- century problem of the p-zombie is that the zombie’s behaviour is different from the human’s. When the p-zombie sees the snake it doesn’t scream and run away, at least not with the speed and computational efficiency of a human. The zombie’s visual system can’t broadcast the snake’s presence to the rest of the mind all at once, for the simple reason that there’s nothing to broadcast it to. Consciousness isn’t purposeless, as Huxley suggested, but we don’t have it because we’re lit with an essence of the divine spark, as the priests and intellectuals he argued with claimed. We’re conscious because it’s more computationally efficient.

When I was wandering the trails at Wangapeka my brain attempted to understand how the maze of forks and branches matched with the partial map of trails in my memory. When I found the lookout I felt happiness: the brain’s way of rewarding the subminds for accurate predictions. If I wound up at a dead end I felt sadness, frustration. The subminds’ models were wrong. Predictive processing sees emotions as responses to events which increase or decrease uncertainty.

But what is mood? Why do we experience sustained feelings of happiness or anxiety or sadness? Why don’t we just exist in a transient stream of pleasure and pain? Friston’s answer is that moods are high-level priors, meaning that they’re predictions about the accuracy of your emotional responses. If we’re wandering through the forest and keep hitting dead ends despite dozens of attempts, we feel anger at each failure and after a while we feel a more persistent sour mood, which is a way of signalling to the subminds that whatever they’re doing, it isn’t working. Then we might succeed, once, and find our way to a branch that looks familiar. But the brain doesn’t want us to feel happy and tell the subminds that their predictive models are correct, because they probably aren’t. If we’ve been wrong dozens of times in a row and right only once, it’s probably just by chance. Bad moods attenuate those signals of happiness.

And if this prior gets stuck in place for a long period of time you’re anhedonic, i.e. you’re never happy, i.e. you’re depressed. Predictive processing suggests that depression is a persistent prior predicting bad outcomes with high confidence. This prior says that we’re always on the wrong path, always heading towards a dead end. When something goes wrong the prior is validated, but when things go right the feelings of happiness or joy are suppressed, because the prior assumes anything good is just statistical noise.

Anxiety happens when the brain predicts bad outcomes with low confidence, and this accounts for that disorder’s classic symptom: obsessional looping. Subminds repeatedly project the same thoughts into consciousness, hoping to reduce error. In the lost-in-the-forest analogy, we think the path ahead simply leads back to where we’re standing right now but we’re not sure, so we take it, but when we come to this spot again we still aren’t sure, so we take it again. And again. And again. Many of the contemporary theories about mood disorders involve genetics, epigenetics, diet, inflammation, environment, stress, early childhood development – really complicated combinations of all of these really complicated things. But there’s one subject that often crops up in the mood disorder literature and that is highly relevant to predictive processing and meaningful to my own experience. Sleep and sleep deprivation.

Sleep has always annoyed biologists. Evolution is supposed to be this ruthless struggle for existence, but every day most complex organisms stop competing for mates and eating each other. Instead they lie down and shut their eyes and ignore everything around them for hours. Sleep should get heavily selected against. We know alternatives to deep sleep are possible. Look at the unihemispheric sleep behaviour of dolphins: the different sides of their brains take turns to sleep so the creatures can keep swimming in the right direction over long distances. And yet deep sleep and dreaming are the norm among complex land animals. Why?

Sleep seems to do a number of important things, and many of them involve computational complexity. If I’m in the maze of trails trying to find my way back to my room, I might pick paths by flipping a coin. I know there’s a 50 per cent chance it’ll come up heads, 50 per cent tails, but even if I don’t know this I could figure it out just by flipping the coin a number of times and observing the results. In other words, I’d build a predictive model. In theory it would be easy to remember this model and make predictions using it, because it’s so simple. 50/50. Heads or tails.

But let’s say there’s some noise in my coin flips and it comes up heads slightly more often than tails. There’s nothing wrong with the coin, it’s just dumb random chance. Existence is full of it. But because my brain is a predictive processor I have to update my model of how the coin behaves, and try to remember there’s a 0.51 probability I’ll flip heads and an 0.49 probability I’ll flip tails. And as I move around I notice something else. If I flip a coin when I’m facing uphill, it’s slightly more likely to come up tails than if I’m facing downhill. Again, this doesn’t mean anything. It’s just more noise, but now I have to update my model again, and it’s getting more complex. And this goes on and on, all day every day. Every prediction and outcome adds more noise to my operating models of the world. It gets harder to store them, and to make predictions based on them, and none of the complication adds any value to my predictive accuracy. It only takes about sixteen hours for my models of reality to become pointlessly complex and expensive. What I need is a process to strip out all that noise.

Every night we sleep and our overly complicated model of the world goes offline. It’s still making predictions but, unless we have a serious sleep disorder, we can’t move our bodies to match up with them, and there’s very little sense data coming in. These predictions are dreams. They seem to ‘prune redundancy and reduce complexity’, as Friston puts it. My hypothetical model for coin flips gets reset to a simple 50/50. And when I’m not suffering from depression, this process resets all those high-level mood priors. If I have a terrible day I sleep on it, and wake feeling fine.

If you meditate for long enough you see and experience a lot of strange things. I’ve had visions, heard voices, felt bizarre bodily sensations. They’re a lot like dreams, but I experience them with total clarity because I’m not asleep. My mind is making predictions, because that’s what it does, but when I’m deep in a meditative state it doesn’t have any external data to match against. I haven’t had an out-of-body experience yet, but I’ve lost all connection with my senses a few times. Both phenomena seem to connect to a brain region called the temporoparietal junction, which integrates sense data, and which is probably doing something odd when people have the common meditative experience that they’ve floated out of their body and into some higher realm of being, or glimpsed another plane of existence.

*

After the jhana, I spent the afternoon walking around the trails. I found some of the trees I liked and just stared at them, delighted that entities of such beauty could exist, manifesting themselves into the world out of nothing. I felt extremely sane. While walking slowly between pines I noticed something in my peripheral vision. It was a weka, a sleek flightless brown- feathered bird, standing about knee height, staring at me from the foliage. We contemplated each other for a few moments, me drinking in the animal’s wet beauty, it probably wondering whether I was likely to chase it down and eat it. I felt sorry for the weka, which would probably end up being killed by a dog, or some other carnivore, against which its evolution on a remote island with no mammalian predators left it defenceless. Then I thought: wasn’t I in a similar position to the weka?

Wasn’t I operating outside my own evolutionary niche? True, I wasn’t in danger of being eaten by dogs. Instead my brain was optimised for low information, high risk, high opportunity environments. But I lived in a leafy suburb and worked at a university, and they were the exact opposite: high information, low risk. I spent every day trying to absorb as much data as I could, reading emails and books and papers, scrolling social media, the news, TV. I did this because my brain thought it was important to form an accurate model of reality and to be alert to risk, like the sleek little weka detecting potential predators. But for me there were no risks and my culture saturated me in an infinite amount of information, which overloaded my brain and made me sick. ‘I am the weka,’ I murmured into the silence of the clearing. The actual weka had already vanished into the ferns.

I thought about all of this again on my beach holiday, reading my neuroscience papers, following trails of footprints across the wet sand. I realised that all of the information I’m drenched with, addicted to, adds noise to my models of reality and generates a huge amount of self-referential chatter between my subminds. This is especially so if I’m spending a lot of time on the internet, interacting with content designed to make me feel anxious or outraged or angry. And, like the dogs from my childhood, pointlessly barking at each other across the valley, this internal chatter has a looping, cumulative effect – information generating information, noise amplifying noise. I forget most of it when I sleep and this gives my life an odd, amnesiac-like quality. Every day I try to learn what’s going on in the world, and every night I forget almost everything.

Friston doesn’t go into a lot of detail in his paper on dreaming. How does it ‘prune and simplify’ our models of reality? He doesn’t say. I am not a neuroscientist, nor an expert on brain imaging, or any kind of expert at all. I’m just a depressive who had a blissful green light explode inside his head.

But the way I imagine it works is that the sleeping mind generates a thought – some memory or fantasy or nightmare – designed to provoke a specific response, then it measures the actual reaction and prunes out all the priors that deviate from the prediction or that make no measurable difference. You have a bad day and go to bed angry, then you have a flying dream and your mind strips out everything that inhibits the predicted response of happiness and elation. And the narrative component of our consciousness strings all these unrelated predictions together, which gives us the weird, nonlinear, non- logical dream narrative, and that’s why hours and hours of uncontrolled hallucinations cause us to wake up feeling normal.

Unless, that is, your brain is so overstimulated, so drowned in information that you wake in the night with your subminds screaming nonsense at each other. If healthy sustained sleep doesn’t happen then none of the pruning takes place; none of the high-level mood priors get reset. So day by day we feel a little more exhausted because we’re trying to make sense of the world with these noisy, computationally expensive models of reality, and our overall mood becomes a sinking lid. It never gets reset. We’re stuck in a depressive state.

For me, this model of depression explains a lot. It explains why insomnia is so closely tied to my depressive episodes. It explains why I want to sit in a dark room and binge-watch the same old TV shows. I’m minimising uncertainty. And it explains why meditation helps treat the disorder. It’s significant, I think, that concentration meditation deals exclusively with simple repetitive subjects: the breath, a mantra, the flame  of a candle. Nothing to complicate the models. I think this disrupts the obsessional looping that’s so characteristic of mood disorders, which keeps those high-level priors stuck on depressive or anxious modes. And, at the same time, I’m training my mind to ignore the internal chatter that feeds it. All day every day my subminds spam my consciousness with noise, and when I meditate I repeatedly ignore them. If I do this often enough – it has to be every day; it doesn’t seem to work if it isn’t – I create a high-level prior of my own: that their endless chatter is not important. My brain doesn’t fill with noise, and the subminds stop waking me at 1am.

I try to meditate for half an hour first thing in the morning, another half an hour in the evening. The morning session seems to be the most effective. But there’s a tension between my desire to meditate because it’s important for my mental health and my desire not to because it’s dull. Over the weeks and months, my practice follows a sawtooth pattern. I’m enthusiastic and diligent for a while, but it’s nice to sleep in on weekends. Then I drop the weeknight evening sessions. I skip a couple of weekly morning sessions. My equilibrium starts to slip, usually so slowly that I don’t notice it until I start waking in the night. There are other subtle and not so subtle signs, like indecision, mood swings, feelings of blind rage if the automatic doors at the supermarket don’t detect my presence. Because I know these are warnings, I rededicate myself to the practice.

There’s a pleasant clarity to the mind when I’m in one  of these dedicated phases. If you think of your subminds as guests at a dinner party, the ones who shout the loudest are the most annoying. If you can get them to calm down you’ll hear from the quieter guests, who tend to have more interesting thoughts. When I’m in this space I also experience moments of no-thought, when I notice something trivial but beautiful in the everyday: most recently the pale green mould growing inside my porch light, revealed when I switch it on at night, and I stop and look at it, and think about nothing.

But it is not a panacea. I can’t meditate my way out of serious depressive episodes. I doubt many other people can either. ‘Try meditation’ is probably not useful advice for people who are very sick, and it might not be useful advice for a lot of people who suffer from anxiety and depression. The brain is complex. Mood disorders are complex. They’re probably multiple conditions with different pathologies that have been grouped together because the symptoms are similar. But I believe that there’s a subset of people like me, whose problems involve information and distraction and uncertainty, and that mindfulness is an effective long-term way to manage our illness. And I find the predictive processing framework useful because it makes sense, at least to me. Which is important, because it’s hard to maintain a daily habit that feels nonsensical even if it works.

It might all be wrong. Psychiatrists and neurologists always think they’re on the verge of understanding the mind and its disorders. They’ve thought this for over a century and all the theories have been wrong. But if our minds are computational, it seems inevitable that some of their problems will be computational: broken loops, intractability, complexity. Noise in the data.

So what are the jhana states? Nobody knows. The meditation author and teacher Leigh Brasington theorises that the first jhana has something to do with distraction. For the purposes of day-to-day survival you don’t want your brain to have reactions to its own reactions. You don’t want to feel pleasure in response to pleasure, and you really don’t want pain in response to pain. So both qualia fade over time. The mind switches its attention to something else.

Economists sometimes talk about ‘Knightian uncertainty’: uncertainty about uncertainty. This is when you don’t know things about the world, but you don’t know what you don’t know, and if you’re an animal like the weka or a pre-modern human, not knowing things can kill you. Rapidly cycling between subminds is an attempt to minimise this risk. But you can circumvent this cycling by taking various recreational drugs that flood the pleasure circuits of your brain with neurotransmitters, or you can use concentration meditation to stay on the sensation of pleasure and allow it to amplify itself, which is how the first jhana seems to work. And this feels very nice. I mean, it is literally bliss. But bliss is less interesting than you think.

If we go back to the Buddha sitting beneath a tree and achieving awakening, whatever that is, the Pāli scriptures teach that he did this first by mastering the jhanas, then by practising a different form of meditation – vipassanā, or insight meditation. This is the technique that allegedly reveals the true nature of reality.

Some people who meditate a lot claim they’ve seen past lives or guardian angels, other planes of existence, and I don’t believe any of them. But I have met dedicated meditators who’ve seen . . . something. They find it hard to describe. Non- existence. Cessation. Emptiness. They say it’s an experience rather than an idea. Whatever they see, it transforms the way they experience reality and this transformation seems to be measurable. All around the world, neuroscientists are stuffing Buddhist monks with hardcore meditation practices into fMRI and PET scanners and finding that their brains look and function differently from everyone else’s. They’re subjecting them to psychological tests and finding that insight meditation makes them calmer and happier than the rest of us.

Concentration meditation is about training your brain to concentrate deeply. Insight meditation is about what you do with that concentration. There are different techniques. Maybe you note every thought or sensation that passes through your mind; maybe you focus intently on the individual components of the breath; or you turn your attention to your body and scan different parts of it to feel, say, your blood pumping through your circulatory system or even just the routine traffic in your peripheral nervous system that your mind normally filters out. When I sit in access concentration for long enough – and the final day of my retreat in Wangapeka, the day after the jhana, is almost the only time I’ve been able to do this – and practise vipassanā techniques, a number of interesting things happen. The first is that I stop thinking in terms of language. I no longer have access to the idea of an inhalation or exhalation, or a nose or an abdomen, or even a breath. That submind no longer projects its thoughts into consciousness. Instead, I experience the sensations of the breath in ways that are a little hard to describe, because the whole experience is wordless, but ‘wavelike’ comes pretty close.

And it’s an unstable state. It’s easy to spill out and think in words again. Eventually it stabilises and I notice there’s a granular quality to my thoughts. Our consciousness seems like an unbroken stream, but during insight meditation it presents itself as a sequence of discrete constructs. Once I observe this I notice the gaps between the thoughts, which we’ve trained ourselves not to notice, like the spaces between words in a sentence. These start out as brief flickers of nothing, but as my thinking simultaneously slows down and speeds up the individual sensations become slower and slower, until I see each separate thought bubble up, out of nothingness, and then evaporate. This is an awe-inspiring experience. When  I think about it afterwards, it feels like I’ve witnessed the heart of the hard problem: the alchemical moment when the electrochemical signals in the brain transform themselves into conscious thought.

People who dedicate years to this type of meditation have insight experiences. They don’t see past lives or heaven. They see – so they say – glimpses of reality, the true nature of mind, the emptiness of material existence. The cumulation of these moments leads to awakening.

I haven’t had many insight experiences. I don’t have time. I have a family, a job, a mental health problem to deal with. Stuff on TV I want to watch. I feel very far from awakened. But I see what they’re saying about emptiness. There is a powerful sense of nullity when I glimpse these moments between thoughts – a sense of annihilation. When I recall it, I think about Heidegger’s premonition of the nothing that lies beyond the visible world. I wonder if depression and anxiety are just biological malfunctions, closed loops, or whether there are alternate ways of seeing and being in the world, which mood disorders point us towards. I think about the time I hid my daughter’s toy and made her cry, and the idea that high-level abstract priors embedded deep in our minds, existing beyond language or rationality – priors like object permanence – lock us into fixed perceptions of reality. I wonder if the insight meditators are updating those priors, if they’re perceiving existence in ways the rest of us cannot.

And I think about trails through the trees, leading to on- ramps and expressways, down unlit halls, footprints on the beach, neurons opening into synaptic clefts, paths through the dunes, recondite pathways forking and branching. I wonder if they all lead back to themselves, a series of interconnecting labyrinths with no exits and no centre, or whether there really is a way through the maze, a clearing in the mind where the world opens up to itself. Some vast neurochemical valley, flooded with probability and light. A place and a state outside the hallucination.