Neuroenergetics

Abstract

The brain uses 20% of the body’s energy. The processes delivering that energy to neurons can fail in numerous ways. The neuroenergetics theory draws out the implications of failure in the supply chain between blood capillaries and neurons. The theory is implemented as a diffusion model that yields response-latency distributions, error rates, and other predictions for typical individuals engaged in focused activities and for special populations such as those with neurodevelopmental disorders. It predicts the effects of stimulants, trial spacing, and fatigue. Here, the implications of energetic insufficiency are explored in the context of the positive manifold of abilities, disabilities, and psychiatric comorbidities.

Keywords

astroglia diffusion-decision models fatigue lactate response time distributions ADHD

Functional MRI (fMRI) has driven recent advances in neuroscience. fMRI measures the increase in blood flow that feeds neurons with oxygen and glucose—the brain’s primary source of energy—in active parts of the brain. It visualizes areas involved in perceptual, cognitive, and motor functions and contrasts functional changes across development with those in neurodevelopmental disorders such as attention-deficit/hyperactivity disorder (ADHD). As functional connectivity increases with the maturation of neuronal networks, so too does the brain’s energetic burden (see, e.g., Vértes & Bullmore, 2015). Information is often routed through network “hubs” to conserve energy. Despite their efficiency, “functional hubs may be highly vulnerable to conditions that endanger their energy supply” (Tomasi, Wang, & Volkow, 2013, p. 13646), a vulnerability that will affect not just a few neurons but much larger functional networks. It is the nature of this energy supply and the implications of its vulnerabilities that motivated the theory sketched here.

The Neuroenergetics Theory (NeT)

Changes in blood flow measured by fMRI reflect changes in neural activity. Decreases in neural activity may occur because a module of the brain is not required or not enlisted, or it may occur because of a bottleneck in the energy supply to that module. Our research leads us to believe that in many cases, decreases in energy supply to neurons do not merely signal hypofunctionality in those regions: They may actually cause that hypofunctionality. This hypothesis runs counter to popular theory, according to which failures of executive function, or of motivation, or tolerance for delay, or of dopamine supply, are held to be a cause of such neurodevelopmental disorders. Failures of the energy supply can occur whenever extended cognitive efforts are required. Over time the brain has evolved many mechanisms to meet the challenges of the moment and has patched them over and over with new mechanisms in a myriad of system updates. That is the case with neural energetics. The following overview therefore must omit many details of these complex systems (for those, see Magistretti & Allaman, 2015; Stobart & Anderson, 2013).

Neurons transmit information in spikes of rapidly changing electrical currents across their membranes. These action potentials travel the length of their axons, releasing neurotransmitters such as glutamate that activate neurons downstream in the network. Attention and memory formation require synchronized high-frequency firing in networks spanning many brain regions (Lord, Expert, Huckins, & Turkheimer, 2013). For that energy, neurons utilize glucose direct from the blood vessels and lactate from surrounding glial cells to make ATP, their ultimate fuel (see Fig. 1). Those glial cells, called astroglia or astrocytes, convert glucose to a fuel more immediately useful to neurons—lactate—and also convert it into and store it as glycogen. Under increased demand, neurons prefer lactate, as that is faster and more efficient to use than glucose, which must be processed from scratch. The glutamate released by neurons when they fire not only signals downstream neurons but also conveniently stimulates astrocytes to acquire glucose from blood vessels to make lactate, which shuttles to neurons for ATP production. This “astrocyte-neuron lactate shuttle” (Magistretti & Allaman, 2015) is one of the key players in the energy supply chain.

Fig. 1.

Schematic representation of glucose and lactate provisioning of neurons. Glucose is the prime energy source of the brain. Fast-spiking neurons prefer to use lactate because it is faster and less costly than the direct use of glucose. Presynaptic neurons release glutamate at synaptic connections to signal downstream neurons; that glutamate also stimulates astrocytes to absorb glucose from blood capillaries and release lactate to neurons for fuel. Norepinephrine released from spiking axons further stimulates astrocytes to convert glucose stored as glycogen to lactate, nourishing the hungry neurons. See Killeen, Russell, and Sergeant (2013) for more detail.

Norepinephrine is another transmitter that stimulates astrocytes to release lactate. Dysregulation of neurons that release norepinephrine may lead to cognitive deficits in conditions such as ADHD (Imeraj et al., 2012). Stimulants used to treat ADHD such as methylphenidate (e.g., Ritalin) and amphetamine (e.g., Adderall) block reuptake of norepinephrine, increasing its extracellular concentration and releasing lactate from astrocytes, restoring some integrity to affected brains (Spencer et al., 2013). All of these facts suggest a key role for energetic insufficiencies in this condition.

As rats age, they become less sensitive to epinephrine and lose their ability to increase blood glucose levels in response to arousal. This decreases the fuel stock for lactate, impairing memory consolidation and retrieval. Administration of glucose slows the rate of forgetting (Gold & Korol, 2014), showing again the key role of arousal-related epinephrine and norepinephrine in the energy budget.

Implications for Behavior

In tasks requiring sustained attention (vigilance, continuous performance, etc.), performance shows decrements with time on task. The fast depletion of energy within the neurons, combined with the slower provisioning from astrocytes’ conversion of glucose and glycogen to lactate (Gundersen, Storm-Mathisen, & Bergersen, 2015), progressively impairs sustained attention and cognition. It takes hours to restore astrocytes’ supplies of glycogen, a task typically completed during sleep. That is why insufficient sleep exacerbates performance decrements, as does mental fatigue generally (Castellanos & Proal, 2012). Stimulants reenergize performance by increasing norepinephrine levels, prompting the release of lactate (Magistretti & Allaman, 2015)—as long as the astrocyte has any glucose or glycogen left available for conversion.

As network hubs deplete their fuel, they become unable to maintain synchronous high-frequency firing across brain regions, which dulls their ability to rekindle memory traces and process information. Responses take longer, their latency becomes more variable, transient targets are missed, and more time is spent off-task; minds wander. Depending on the task demands and energy supply, performance eventually stabilizes at lower levels.

Diffusion Models

A class of models called diffusion-decision (or sometimes drift-diffusion) models (DDMs) can translate the above physiological model into one with interpretable parameters and testable predictions. The bottom panel of Figure 2 charts the hypothetical accumulation of evidence, pro and con, toward a threshold of decision. The rugged appearance of these “random walks with drift” reflects the random ebb and flow of excitation and inhibition. Each proceeds at the same average drift rate toward the decision threshold, or criterion. The time required to cross the threshold, over many trials, is traced by the smooth curve above it, the response-time (RT) distribution. Simen, Rivest, Ludvig, Balci, and Killeen (2013) have provided a history of such models, and Ratcliff, Smith, and McKoon (2015) their current status.

Fig. 2.

A schematic illustration of the drift-decision model of response times. Panel (a) depicts a response-time distribution that could arise from the accumulation of information to a criterion sufficient for a response. Panel (b) shows eight random walks with drift, representing the interplay of neural processes of excitation and inhibition that cumulate in a response when the criterion at the top of the panel (75 steps in the positive direction) is crossed. Simulation of the process many times results in the distribution of criterion crossings shown in panel (a). Adapted from “Absent Without Leave; A Neuroenergetic Theory of Mind Wandering,” by P. R. Killeen, 2013, Frontiers in Psychology, 4, Article 373. Copyright 2013 by the author.

In the simplest case, such as that pictured in Figure 2, the RT distribution is the Wald density. Because few studies in the literature had conducted DDM analyses, Killeen, Russell, and Sergeant (2013) estimated the crucial measures of drift speed and criterion from the published parameters. This was also the tactic of Huang-Pollock, Karalunas, Tam, and Moore (2012) in a meta-analysis showing that speed of information processing—drift speed—is the major difference between ADHD and control groups. They also found, consistent with the predictions of NeT (Killeen et al., 2013, pp. 647–648; Russell et al., 2006), that average processing speed was faster for shorter sessions; it decreased as a function of session length and was faster for slower-paced stimulus presentation, which provided time for the repletion of energetic stores.

Application of the DDM to data of individual subjects permits estimation of sensory encoding and motor execution speed, information-processing speed (drift rate), and speed–accuracy trade-off (criterion placement) that inform any decision process. Because it takes these variables into simultaneous consideration, it avoids confounds in which changes in drift rate might be compensated by changes in criterion setting. Many variables can affect drift rate; as individuals learn a task, for instance, processing becomes more efficient: Drift rates increase. At the same time, there will be slowing on a different time scale due to energetics (Killeen et al., 2013). Enough studies using DDM have now been conducted to support a meta-analysis of the results (Karalunas, Geurts, Konrad, Bender, & Nigg, 2014) testing the central prediction of NeT. Information-processing speed—drift rates—will be slower and response times more variable in subjects with neuroenergetic compromise: Hungry neurons cannot maintain the firing speed necessary for optimal computations (Lord et al., 2013). Increased variability is a natural concomitant of slower neural processing, since the variance of the Wald distribution increases as the reciprocal of the cube of the drift speed (v): σ² = c/v³. The meta-analysis captured six studies involving 500 children and adolescents with ADHD and 600 typically developing controls in a variety of tasks. Every one of the studies showed a significant group difference on drift speeds, slower for those with ADHD than controls, with a large overall effect size of g = 0.63. There was no systematic effect on criterion (“boundary separation”), but controls were slower at motor execution (g = 0.32).

A recent study (Huang-Pollock et al., 2016) tested another prediction of NeT, one concerning event rate (ER, which varies inversely with intertrial interval, ITI). If individuals must remain alert during the ITI (as is the case when the time of onset of the next stimulus is uncertain), low ER will fatigue the brain more for a fixed number of trials (but not for a fixed session length) than high ER. If ER is manipulated using signaled ITIs, it gives the brain a brief time to “go off-line” and conserve, even replenish, energetic resources. Huang-Pollock and associates varied ER by extending the stimulus screen after a response had been made and adding a brief signaled ITI to generate a low-ER condition. As predicted by NeT, drift speeds were faster in the low-ER condition. The authors suggested that earlier studies that found different effects of ER did not use a DDM and therefore were subject to the potential confound of compensatory changes in criterion.

Newman, Korol, and Gold (2011) studied memory formation and retrieval in rats over extended sessions. They found that lactate derived from conversion of glucose or glycogen by astrocytes “may be an important substrate for neurons during working memory by providing rapid additional energy at times of high need. As shown [in their report], that need can be generated by cognitive demands” (p. 8). Blocking the processing of glycogen impaired memory; infusing lactate into the brain improved it. Given the time constants of these processes, their work provides the neural mechanism for the ER found with humans. It also points toward the role of neuroenergetics in working memory.

Extending the NeT

NeT should predict performance on any task involving drift-decision processes. Such processes are a plausible mechanism for interval timing (Simen et al., 2013). The coefficient of variation of threshold crossings in the Wald distribution varies inversely with drift rate. NeT therefore predicts reductions in relative timing accuracy (i.e., Weber fractions) when functional brain networks fatigue or in special conditions such as ADHD. This has been observed (Toplak, Dockstader, & Tannock, 2006).

On continuous performance and other boring (i.e., focal-energy-depleting) tasks, there is often a change of phase—the mind wanders. The extension of NeT to this case (Killeen, 2013; Killeen et al., 2013) delivers response distributions that are a mixture of Wald latencies (from the attentive state) and ex-Wald latencies (from the fugue state). Similar effects are achieved by allowing the drift rate to vary, a feature of many DDMs.

The parameters of NeT correlate with independent measures of the constructs. Stimulants increase drift rate in cognitive tasks, and time on task decreases it (Killeen et al., 2013; see Figs. 12, 6, and 7). A study reporting performance on a perceptual-motor task and a scale of impulsivity (Logan, Schachar, & Tannock, 1997) revealed a correlation of r = −.84 between impulsivity and the criterion, as inferred from the authors’ RT parameters.

Stimulants help maintain attention because they energize the brain as a whole in myriad ways (O’Donnell, Zeppenfeld, McConnell, Pena, & Nedergaard, 2012). They provide a boost to a brain that is struggling to keep up with the demands placed on it. Administered therapeutically, they are in fact more prosthetic than therapeutic. Different types of interventions help individuals with challenges at different points of the supply chain: iron therapy for those with low stores of iron—which includes the majority of children with ADHD (Konofal et al., 2008); fish oil (Sonuga-Barke et al., 2014); micronutrients (Kaplan, Rucklidge, Romijn, & McLeod, 2015); good sleep discipline (Cortese, Faraone, Konofal, & Lecendreux, 2009); and vigorous play and exercise for all (Halperin & Healey, 2011). Each contributes to brain energetics and may alleviate some of the symptoms that all of us are heir to, and some of us have too richly inherited.

Distributed Systems

The brain is a complex nexus of hubs and networks with no single part generally identifiable as the “center of x” (Sporns, 2013; Uttal, 2015). The mind is likewise divided (Van Orden & Paap, 1997): Godfrey Thomson formulated a distributed processing model of intellectual ability a century ago (Bartholomew, Deary, & Lawn, 2009), echoed in recent discussions of the “positive manifold” (Anderson, 2010; Van Der Maas et al., 2006) and “neural reuse” (Rabaglia, Marcus, & Lane, 2011). For those with intellectual disabilities, the positive manifold is an imperfect one of subtle distributed inefficiencies. Most disorders in the Diagnostic and Statistical Manual of Mental Disorders (DSM) are heterogeneous, with overlapping symptoms showing no “zones of rarity” among them (Lilienfeld, 2014). Perhaps this is inevitable, given the many small interactions of many genes that underlie abilities and disabilities (Chabris, Lee, Cesarini, Benjamin, & Laibson, 2015). Given the stochastic nature of early brain connections, the neuroenergetic insufficiencies hypothesized by NeT will inevitably have different impacts on different hubs, causing corresponding signs of various, often comorbid disabilities that may cumulate into one diagnostic category or another—or simply contribute to an individual’s character. Fried (2015) has argued that, because the symptoms of depression do not constitute a syndrome (i.e., a taxon), it is more profitable to study and treat the signs and symptoms. The same may be said of most categories in the DSM. The novel index of drift rate may map informatively onto one of the principal components of this sign-and-symptom space, contributing to a new dimensional nosology (Karalunas et al., 2014; Lahey & Willcutt, 2010).

There are many ways in which the delicate and complex energetic supply to the brain can falter. It does so regularly for anyone who is tired, is undernourished, or has focused on a task for too long. Special vulnerabilities have many causes: Late maturation of the insulation of neurons, insensitivity to norepinephrine, failures of mitochondria to create adequate ATP, or failures of other components of the energy supply chain. Such vulnerabilities may even have an evolutionary advantage in harsh environments (Killeen, Tannock, & Sagvolden, 2012) and, once triggered by early stress, become a part of character.

These are some of the current directions in research on brain energetics in typical and atypical development, but they are directions on a compass with no true north. There are many directions to explore, the neuroenergetics hypothesis among them. No simple cause such as “inhibitory failure” or “executive failure” will be found for conditions such as ADHD. Instead, there will be a broad spectrum of causes, and of relevant interventions, one of which may help one or another person, whether he or she is categorized as atypical or is just another member of the species dealing with the imperfections of the human condition.

Footnotes

Author Contributions

V. A. Russell organized the neuroscience, R. Tannock the clinical implications, and P. R. Killeen and V. A. Russell the models and manuscript.

Declaration of Conflicting Interests

The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.

Funding

R. Tannock has received speaker fees and travel costs from Shire, Lilly, and Pearson-Cogmed, as well as software from Pearson-Cogmed, for her nationally funded research on working memory training. She was a member of the DSM-5 Work Group on ADHD and Disruptive Behavior Disorders.

References

Arnsten

A. F. T.

Rubia

(2012). Neurobiological circuits regulating attention, cognitive control, motivation, and emotion: Disruptions in neurodevelopmental psychiatric disorders. Journal of the American Academy of Child & Adolescent Psychiatry, 51, 356–367. An article that reviews basic and clinical evidence concerning the role of neural networks of the prefrontal cortex in behavior and cognition that are impaired in neurodevelopmental disorders.

Killeen

P. R.

Russell

V. A.

Sergeant

J. A.

(2013). (See References). Contains most of the refinements, applications and predictions of the neuroenergetics theory.

Lord

L.-D.

Expert

Huckins

J. F.

Turkheimer

(2013). (See References). Discusses the energy requirements of synchronized high-frequency excitatory and inhibitory neurons in distributed neural networks that support the functional connectivity of brain regions.

Magistretti

P. J.

Allaman

(2015). (See References). A cellular perspective on brain energy metabolism and functional imaging. Neuron, 86, 883–901. Provides a comprehensive overview of the cellular and molecular basis of the astrocyte-neuron lactate shuttle.

Ratcliff

Smith

McKoon (2015). (See References). Provides a synopsis of current drift-decision models, and in particular the Ratcliff-McKoon research program.

Anderson

M. L.

(2010). Neural reuse: A fundamental organizational principle of the brain. Behavioral & Brain Sciences, 33, 245–266.

Bartholomew

D. J.

Deary

I. J.

Lawn

(2009). A new lease of life for Thomson’s bonds model of intelligence. Psychological Review, 116, 567–579.

Castellanos

F. X.

Proal

(2012). Large-scale brain systems in ADHD: Beyond the prefrontal–striatal model. Trends in Cognitive Sciences, 16, 17–26.

Chabris

C. F.

Lee

J. J.

Cesarini

Benjamin

D. J.

Laibson

D. I.

(2015). The fourth law of behavior genetics. Current Directions in Psychological Science, 24, 304–312.

10.

Cortese

Faraone

S. V.

Konofal

Lecendreux

(2009). Sleep in children with attention-deficit/hyperactivity disorder: Meta-analysis of subjective and objective studies. Journal of the American Academy of Child & Adolescent Psychiatry, 48, 894–908.

11.

Fried

E. I.

(2015). Problematic assumptions have slowed down depression research: Why symptoms, not syndromes are the way forward. Frontiers in Psychology, 6, Article 309. doi:10.3389/fpsyg.2015.00309

12.

Gold

P. E.

Korol

D. L.

(2014). Forgetfulness during aging: An integrated biology. Neurobiology of Learning and Memory, 112, 130–138.

13.

Gundersen

Storm-Mathisen

Bergersen

L. H.

(2015). Neuroglial transmission. Physiological Reviews, 95, 695–726.

14.

Halperin

J. M.

Healey

D. M.

(2011). The influences of environmental enrichment, cognitive enhancement, and physical exercise on brain development: Can we alter the developmental trajectory of ADHD? Neuroscience & Biobehavioral Reviews, 35, 621–634.

15.

Huang-Pollock

C. L.

Karalunas

S. L.

Tam

Moore

A. N.

(2012). Evaluating vigilance deficits in ADHD: A meta-analysis of CPT performance. Journal of Abnormal Psychology, 121, 360–371.

16.

Huang-Pollock

C. L.

Ratcliff

McKoon

Shapiro

Weigard

Galloway-Long

(2016). Using the diffusion decision model to explain cognitive deficits in attention deficit hyperactivity disorder. Manuscript submitted for publication.

17.

Imeraj

Sonuga-Barke

Antrop

Roeyers

Wiersema

Bal

Deboutte

(2012). Altered circadian profiles in attention deficit/hyperactivity disorder: An integrative review and theoretical framework for future studies. Neuroscience & Biobehavioral Reviews, 36, 1897–1919.

18.

Kaplan

B. J.

Rucklidge

J. J.

Romijn

McLeod

(2015). The emerging field of nutritional mental health: Inflammation, the microbiome, oxidative stress, and mitochondrial function. Clinical Psychological Science, 3, 964–980.

19.

Karalunas

S. L.

Geurts

H. M.

Konrad

Bender

Nigg

J. T.

(2014). Annual Research Review: Reaction time variability in ADHD and autism spectrum disorders: Measurement and mechanisms of a proposed trans-diagnostic phenotype. Journal of Child Psychology and Psychiatry, 55, 685–710.

20.

Killeen

P. R.

(2013). Absent without leave; a neuroenergetic theory of mind wandering. Frontiers in Psychology, 4, Article 373. doi:10.3389/fpsyg.2013.00373

21.

Killeen

P. R.

Russell

V. A.

Sergeant

J. A.

(2013). A behavioral neuroenergetics theory of ADHD. Neuroscience & Biobehavioral Reviews, 37, 625–657.

22.

Killeen

P. R.

Tannock

Sagvolden

(2012). The four causes of ADHD: A framework. In Stanford

S. C.

Tannock

(Eds.), Behavioral neuroscience of attention deficit hyperactivity disorder and its treatment (Vol. 9, pp. 391–425). Berlin, Germany: Springer-Verlag.

23.

Konofal

Lecendreux

Deron

Marchand

Cortese

Zaïm

. . . Arnulf

(2008). Effects of iron supplementation on attention deficit hyperactivity disorder in children. Pediatric Neurology, 38, 20–26.

24.

Lahey

B. B.

Willcutt

E. G.

(2010). Predictive validity of a continuous alternative to nominal subtypes of attention-deficit/hyperactivity disorder for DSM–V. Journal of Clinical Child & Adolescent Psychology, 39, 761–775.

25.

Lilienfeld

S. O.

(2014). DSM-5: Centripetal scientific and centrifugal antiscientific forces. Clinical Psychology: Science and Practice, 21, 269–279.

26.

Logan

G. D.

Schachar

R. J.

Tannock

(1997). Impulsivity and inhibitory control. Psychological Science, 8, 60–64.

27.

Lord

L.-D.

Expert

Huckins

J. F.

Turkheimer

F. E.

(2013). Cerebral energy metabolism and the brain’s functional network architecture: An integrative review. Journal of Cerebral Blood Flow & Metabolism, 33, 1347–1354.

28.

Magistretti

P. J.

Allaman

(2015). A cellular perspective on brain energy metabolism and functional imaging. Neuron, 86, 883–901.

29.

Newman

L. A.

Korol

D. L.

Gold

P. E.

(2011). Lactate produced by glycogenolysis in astrocytes regulates memory processing. PLoS ONE, 6(12), Article e28427.

30.

O’Donnell

Zeppenfeld

McConnell

Pena

Nedergaard

(2012). Norepinephrine: A neuromodulator that boosts the function of multiple cell types to optimize CNS performance. Neurochemical Research, 37, 2496–2512.

31.

Rabaglia

C. D.

Marcus

G. F.

Lane

S. P.

(2011). What can individual differences tell us about the specialization of function? Cognitive Neuropsychology, 28, 288–303.

32.

Ratcliff

Smith

P. L.

McKoon

(2015).Modeling regularities in response time and accuracy data with the diffusion model. Current Directions in Psychological Science, 24, 458–470.

33.

Russell

V. A.

Oades

R. D.

Tannock

Killeen

P. R.

Auerbach

J. G.

Johansen

E. B.

Sagvolden

(2006). Response variability in attention-deficit/hyperactivity disorder: A neuronal and glial energetics hypothesis. Behavioral and Brain Functions, 2, 25.

34.

Simen

Rivest

Ludvig

E. A.

Balci

Killeen

P. R.

(2013). Timescale invariance in the pacemaker-accumulator family of timing models. Timing & Time Perception, 1, 159–188.

35.

Sonuga-Barke

E. J.

Brandeis

Cortese

Daley

Ferrin

Holtmann

, . . . European ADHD Guidelines Group (2014). Nonpharmacological interventions for ADHD: Systematic review and meta-analyses of randomized controlled trials of dietary and psychological treatments. American Journal of Psychiatry, 170, 275–289.

36.

Spencer

T. J.

Brown

Seidman

L. J.

Valera

E. M.

Makris

Lomedico

. . . Biederman

(2013). Effect of psychostimulants on brain structure and function in ADHD: A qualitative literature review of magnetic resonance imaging-based neuroimaging studies. The Journal of Clinical Psychiatry, 74, 902–917.

37.

Sporns

(2013). Structure and function of complex brain networks. Dialogues in Clinical Neuroscience, 15, 247–262.

38.

Stobart

J. L.

Anderson

C. M.

(2013). Multifunctional role of astrocytes as gatekeepers of neuronal energy supply. Frontiers in Cellular Neuroscience, 7, Article 38. doi:10.3389/fncel.2013.00038

39.

Tomasi

Wang

G.-J.

Volkow

N. D.

(2013). Energetic cost of brain functional connectivity. Proceedings of the National Academy of Sciences, USA, 110, 13642–13647.

40.

Toplak

M. E.

Dockstader

Tannock

(2006). Temporal information processing in ADHD: Findings to date and new methods. Journal of Neuroscience Methods, 151, 15–29.

41.

Uttal

W. R.

(2015). Macroneural theories in cognitive neuroscience. New York, NY: Psychology Press.

42.

Van Der Maas

H. L.

Dolan

C. V.

Grasman

R. P.

Wicherts

J. M.

Huizenga

H. M.

Raijmakers

M. E.

(2006). A dynamical model of general intelligence: The positive manifold of intelligence by mutualism. Psychological Review, 113, 842–861.

43.

Van Orden

G. C.

Paap

K. R.

(1997). Functional neuroimages fail to discover pieces of mind in the parts of the brain. Philosophy of Science, 64, S85–S94.

44.

Vértes

P. E.

Bullmore

E. T.

(2015). Growth connectomics: The organization and re-organization of brain networks during normal and abnormal development. Journal of Child Psychology and Psychiatry, 56, 299–320.