What Neuroplasticity Actually Means
Neuroplasticity is one of the most over-used and under-defined terms in popular neuroscience. In its precise scientific usage, neuroplasticity refers to the brain's capacity to reorganize its structure, function, and connectivity in response to experience, learning, injury, or environmental demand. It is not a single phenomenon — it encompasses at least three distinct biological processes operating at different scales and timescales.
Synaptic plasticity operates at the level of individual synaptic connections, changing how strongly one neuron activates another. Long-term potentiation (LTP) and long-term depression (LTD) are the primary molecular mechanisms. Structural plasticity involves physical changes in neuronal morphology — the growth or pruning of dendritic spines (the postsynaptic contact points on dendrites), changes in axonal arborization, and the formation or elimination of entire synapses. Adult neurogenesis — the generation of entirely new neurons in the adult brain — occurs primarily in the hippocampal dentate gyrus and the olfactory bulb, and represents the most dramatic form of structural remodelling.
All three processes are activity-dependent: they require not just experience but the right kind of experience — novel, challenging, attentive engagement with the world, not passive stimulation. This distinction has profound implications for brain training.
The Cellular Mechanism: Long-Term Potentiation
Long-term potentiation (LTP), first described by Tim Bliss and Terje Lømo in 1973, is the persistent strengthening of a synaptic connection following its repeated activation. The key molecular player is the NMDA receptor — an ionotropic glutamate receptor that acts as a coincidence detector: it opens only when two conditions are simultaneously met. Glutamate must be released from the presynaptic neuron (signalling that the presynaptic cell fired), and the postsynaptic membrane must already be partially depolarised by prior activity (signalling that the postsynaptic cell is also active).
When both conditions are met, calcium enters the postsynaptic neuron through the NMDA channel, activating a kinase cascade (particularly CaMKII — calcium/calmodulin-dependent protein kinase II) that phosphorylates and inserts additional AMPA receptors into the postsynaptic density. More AMPA receptors mean a stronger response to any future glutamate release at that synapse — the synapse is potentiated. This is the cellular embodiment of Hebb's rule: "neurons that fire together wire together." The synapse is now more likely to pass signals from the presynaptic to the postsynaptic neuron in the future, encoding the association.
BDNF (brain-derived neurotrophic factor) is required for the maintenance phase of LTP, sometimes called late-phase LTP or L-LTP. Blocking protein synthesis prevents L-LTP without affecting the initial potentiation, demonstrating that structural synaptic consolidation — involving new protein production and physical remodelling of dendritic spines — underlies the persistence of long-term memories.
BDNF: The Master Regulator of Brain Plasticity
Brain-derived neurotrophic factor is a small signalling protein in the neurotrophin family that supports the survival, growth, and differentiation of neurons throughout life. BDNF binds to the TrkB receptor on postsynaptic neurons and activates three major downstream signalling pathways: MAPK/ERK (promoting neuronal differentiation and long-term survival), PI3K/Akt (suppressing apoptosis and supporting metabolic function), and PLCγ (driving synaptic strengthening). It is the primary molecular signal driving adult hippocampal neurogenesis — the new neurons born in the dentate gyrus that appear critical for pattern separation in episodic memory.
The "BDNF hypothesis of depression" (Duman and Bhagya, 2008) proposed that chronic stress reduces hippocampal BDNF, causing the dendritic atrophy and neurogenesis suppression seen in depression and stress-related conditions, while antidepressants and exercise increase BDNF and reverse these changes. This hypothesis has been substantially supported by subsequent animal and human evidence — though the relationship is bidirectional and the causation is complex.
From a practical standpoint, the most potent BDNF-increasing interventions are:
- Aerobic exercise — produces the largest and most replicable acute BDNF increases, with chronic training elevating basal levels. Running is particularly effective due to the rhythmic motor pattern activating endocannabinoid-BDNF signalling.
- Intermittent fasting and caloric restriction — ketone body β-hydroxybutyrate acts as an HDAC inhibitor that upregulates BDNF gene expression in the hippocampus.
- Cognitive novelty — engaging with genuinely challenging new learning activates cholinergic and noradrenergic projections to the hippocampus that drive BDNF release.
- DHA (omega-3) — docosahexaenoic acid supplementation increases BDNF in clinical trials, particularly in populations with low baseline intake.
- Sleep — BDNF-mediated synaptic consolidation occurs primarily during slow-wave sleep; sleep restriction reduces BDNF expression.
What the BDNF mechanism reveals about the evidence hierarchy of neuroplasticity interventions is that any intervention reliably shown to increase BDNF has a strong a priori case for neuroplasticity benefits — and this is precisely why aerobic exercise sits so consistently at the top of evidence rankings. Erickson's 2011 PNAS study demonstrated that one year of aerobic exercise increased hippocampal volume by 2% and serum BDNF simultaneously, providing direct mechanistic and structural evidence in a single RCT. Isolated cognitive training apps show much weaker effects despite direct brain stimulation because they generate comparatively smaller BDNF responses — the cellular signal for lasting structural change is simply not as potent. The practical takeaway from the BDNF evidence hierarchy is that systemic interventions — aerobic exercise, quality sleep, social engagement, and learning genuinely novel complex skills — are the primary drivers of BDNF elevation and therefore of durable neuroplasticity, with direct brain training as a secondary reinforcer that modifies existing circuitry but rarely drives the same magnitude of new structural growth.
The Lumosity Lawsuit: What the Science Actually Showed
In January 2016, the US Federal Trade Commission settled with Lumos Labs (maker of Lumosity) for $2 million on charges of deceptive advertising. The company had claimed its app could help users perform better at work and school and protect against age-related cognitive decline — claims the FTC found unsubstantiated. This followed a 2014 open letter signed by 75 neuroscientists and cognitive psychologists challenging far-transfer claims from commercial brain training.
The core scientific issue is the distinction between near transfer and far transfer. Near transfer is uncontroversial: if you practice Lumosity's memory game, you get better at that specific memory game and very similar tasks. Far transfer — the claim that getting better at a game-like task improves genuinely different real-world cognitive abilities — has consistently weak and inconsistently replicated evidence.
The most comprehensive meta-analysis, Melby-Lervåg and Hulme (Psychological Bulletin, 2013, updated 2016), examined 87 working-memory training studies and found reliable near-transfer to untrained working memory tasks, but no significant far transfer to reading comprehension, arithmetic, fluid intelligence, or daily life functioning. This finding has been replicated by multiple subsequent meta-analyses. The 2014 ACTIVE study — the largest and longest RCT of cognitive training (2,832 participants, 10-year follow-up) — found that speed-of-processing training reduced dementia incidence by 29% at 10 years, but this was a single specific intervention studied under rigorous conditions, not a generalised "brain training" endorsement.
What Actually Works: The Evidence-Ranked Approaches
When ranked by quality of evidence for real-world cognitive benefit (not just improvement on trained tasks), the hierarchy is clear:
| Intervention | Evidence Level | Primary Mechanism | Effect Size |
|---|---|---|---|
| Aerobic exercise | Strong RCT evidence | BDNF, hippocampal neurogenesis, cerebrovascular | Moderate–large (executive function) |
| Learning a new language | Observational + structural MRI | White matter remodelling, bilingual executive control | Large structural change |
| Musical training | Observational + RCTs in children | Cortical map expansion, auditory processing | Moderate–large (auditory, motor) |
| Dual n-back training | Moderate RCT evidence | Working memory capacity, fluid intelligence | Small–moderate far transfer |
| Commercial brain training games | Near transfer only | Task-specific skill acquisition | Near transfer: large; far: minimal |
The meta-principle from the rankings is that the interventions with the strongest effect sizes — aerobic exercise, sleep optimization, social engagement, learning novel complex skills — are all lifestyle-based and available at zero direct cost. Pharmacological cognitive enhancers like modafinil or prescription stimulants produce acute performance effects, but the evidence for sustained neuroplasticity enhancement from these substances is weak, and their mechanisms do not produce the BDNF-driven structural changes that underlie durable cognitive improvement. The practical protocol derived from the evidence table: prioritize 150 or more minutes per week of aerobic exercise as the primary intervention (largest effect, strongest mechanistic evidence), optimize sleep to 7–9 hours nightly (required for BDNF-mediated consolidation of any learning), and deliberately learn a cognitively demanding novel skill — a second language, a musical instrument, chess — that keeps you operating at the edge of your competence. Layer in evidence-based supplements such as omega-3 DHA (for BDNF support in those with low dietary intake) and B vitamins (for those with identified deficiency) as adjuncts to this foundation, not as replacements for it.
Aerobic Exercise: The Gold Standard
Aerobic exercise occupies a unique position in the neuroplasticity evidence base because it produces demonstrable far transfer to cognitive domains that were never trained. The pivotal RCT was Kirk Erickson's 2011 study (PNAS) in 120 sedentary older adults randomised to either aerobic walking or stretching for 1 year. MRI showed that aerobic exercise increased hippocampal volume by 2% — reversing approximately 1–2 years of age-related atrophy — while the stretching group showed the expected 1.4% decline. The increase in hippocampal volume correlated significantly with improved spatial memory and with increased serum BDNF levels, providing a direct mechanistic link.
The broader meta-analysis by Colcombe and Kramer (Psychological Science, 2003) of 18 RCTs found aerobic training improved cognitive function in all measured domains, with the largest effects in tasks requiring executive function — the set of higher-order cognitive processes (working memory, planning, cognitive flexibility, inhibitory control) subserved by prefrontal cortex and its connections. Cardiovascular fitness improvement, BDNF elevation, increased cerebral blood flow, and reduced systemic inflammation all contribute to this effect.
Sleep: The Non-Negotiable Consolidation Phase
Any discussion of neuroplasticity and learning is incomplete without addressing sleep. Matthew Walker's laboratory and others have established that slow-wave sleep (stages N2 and N3) is required for hippocampus-to-neocortex memory transfer — the offline consolidation process in which newly encoded episodic memories are replayed in the hippocampus and gradually integrated into stable cortical representations.
BDNF-mediated synaptic strengthening — the persistence phase of LTP — occurs predominantly during sleep. REM sleep plays a role in synaptic homeostasis (the pruning of weakened or redundant connections) that is equally important for cognitive performance as potentiation. Chronic sleep restriction to 6 hours per night produces progressive cognitive impairment across all domains — impairment that subjects systematically underestimate. Critically, this cognitive impairment is not fully reversed by a single recovery night.
The practical implication: any brain training, learning, or cognitive intervention is operating at a fraction of its potential when paired with insufficient sleep. The consolidation window that sleep provides is not optional — it is where the plastic changes initiated during wakefulness are stabilised into lasting memory.