There's a ~25% chance you won't use inner speech to read this article on inner speech

Inner speech is the internal experience of language without producing sound or moving your mouth. It differs from silent speech or subvocalization, which involves attempting to speak without sound and activates speech motor systems. True inner speech is purely imagined language: thinking in words without any articulatory effort.

About 25% of people report experiencing little to no inner speech in their daily thinking. For those who do experience it, inner speech serves multiple cognitive functions: maintaining information in working memory, regulating behavior during difficult tasks, and supporting reading comprehension. Recent technological advances have demonstrated that inner speech generates robust neural signals in motor cortex that can be decoded in real-time, enabling paralyzed individuals to communicate by imagining words (Kunz et al., 2025). Meanwhile, clinical research reveals that inner speech becomes disrupted in predictable ways across conditions including schizophrenia, aphasia, and autism, while remaining remarkably variable in typical populations both in frequency (ranging from 0% to 75% of sampled moments) and phenomenology (from condensed fragments to elaborate internal dialogues).

This synthesis examines how inner speech develops, what neural mechanisms support it, which cognitive functions depend on it, and how it varies across individuals, languages, and clinical populations.

Cognitive functions

Inner speech serves multiple cognitive functions, though debates continue about whether it is necessary or merely beneficial for specific tasks. Vygotsky's theory (Vygotsky, 1934) proposed that inner speech emerges through internalization of social dialogue, with its primary function being self-regulation - the verbal mediation of one's own behavior and cognition. Contemporary research has identified additional functions: cognitive scaffolding (organizing complex thoughts and maintaining goals), metacognition (reflecting on one's own mental states), and social cognition (facilitating perspective-taking through internalized dialogue).

Working memory and the phonological loop

The relationship between inner speech and working memory represents one of the most extensively studied domains. Baddeley and Hitch proposed a phonological loop (Baddeley & Hitch, 1974) - a system for temporarily storing and rehearsing verbal information. The model has two components: a phonological store (passive acoustic/phonological storage with 1-2 second decay) and an articulatory rehearsal process (active maintenance through subvocal repetition, essentially inner speech).

Evidence supports inner speech's role in working memory. First, articulatory suppression (preventing inner speech by having participants repeat irrelevant words aloud) reliably impairs memory for verbal material. Second, word length effects show that longer words are harder to remember, suggesting time-based decay during rehearsal. Third, phonological similarity effects indicate that information is stored in phonological form - words that sound alike (like "cat," "mat," "bat") are more easily confused than words with different sounds.

While inner speech and the phonological loop share overlapping neural substrates (left inferior frontal gyrus, supplementary motor area, posterior temporal regions) and similar developmental timelines (both becoming robust around age 7), important distinctions exist: inner speech has functions beyond rehearsal, phonological loop capabilities can be demonstrated before inner speech fully develops, and inner speech phenomenology is richer than phonological coding alone. The flexible abstraction hypothesis (Oppenheim & Dell, 2010) suggests inner speech can vary from abstract (phoneme-level representation) to concrete (full phonological features) depending on task demands and degree of articulation. A 2025 critical analysis by Hughes in the Quarterly Journal of Experimental Psychology (Hughes, 2025) even questioned whether a distinct short-term phonological store is necessary, proposing alternative perceptual-motor approaches.

Consensus holds that inner speech is strongly associated with the phonological loop's articulatory rehearsal component but they are not identical constructs. Inner speech encompasses broader self-regulatory and metacognitive functions beyond memory maintenance.

Executive functions show selective dependencies

Inner speech supports task switching and cognitive flexibility. Articulatory suppression consistently increases "switch costs" - the time and error penalties when alternating between task rules (Emerson & Miyake, 2003) (Miyake et al., 2004). Inner speech appears to function as an internal cue that helps maintain task sets and prepare for rule transitions, with effects most pronounced when external cues are minimal or ambiguous.

Furthermore, inner speech aids self-control and inhibition, particularly in Go/NoGo paradigms. In a typical Go/NoGo task, participants see a rapid sequence of stimuli (like letters or shapes) and must press a button for "Go" stimuli (e.g., the letter X) but withhold their response for "NoGo" stimuli (e.g., the letter Y). Inner speech aids performance by verbally mediating the inhibition process. When facing a NoGo stimulus, participants may internally say "stop" or "don't press" to override their prepotent response tendency. Studies using articulatory suppression (having participants repeat irrelevant words aloud, which blocks inner speech) show that preventing this verbal self-instruction significantly increases NoGo errors - people press the button when they shouldn't. The effect is particularly strong when the task requires quick responses and when Go trials are more frequent than NoGo trials, creating a strong response habit that must be inhibited. Inner speech essentially provides an internal verbal "brake" that helps pause automatic motor responses.

For planning tasks, evidence is mixed. The Tower of London is a classic planning task where participants see two displays: a starting configuration and a goal configuration, each showing three colored beads stacked on three pegs of different heights. The challenge is to mentally plan how to transform the starting arrangement into the goal arrangement in the minimum number of moves, following strict rules: move only one bead at a time, and each peg has a limited capacity based on its height.

Children show clear benefits of inner speech in Tower of London performance. For adults, some studies find that interfering with inner speech disrupts performance beyond general dual-task effects, with these disruptions correlating with real-world executive functioning (Wallace et al., 2009). However, other studies find no articulatory suppression effects on planning. The likely explanation is that planning tasks are fundamentally visuospatial, so verbal strategies are supplementary rather than necessary. Benefits depend on task demands and individual preferences.

Self-regulation consistently links to inner speech across development and domains. Children with better-developed inner speech show superior performance on behavioral inhibition and delayed gratification tasks. In sports psychology, self-talk enhances performance, confidence, and motivation. Interestingly, interrogative self-talk ("Will I?") produces better task engagement than declarative self-talk ("I will") (Senay et al., 2010) (Dolcos & Albarracín, 2014).

Reading comprehension requires inner speech for difficult texts

The relationship between inner speech and reading is particularly well-established. Multiple lines of evidence demonstrate that silent reading activates inner speech mechanisms. A 2011 study used eye-tracking during silent reading of limericks and found disruption when rhyme patterns mismatched readers' regional accents (Filik & Barber, 2011) (Northern versus Southern English), indicating that inner speech reflects personal voice characteristics. Responses are slower for phonetically long words versus short words during silent reading even when orthographically matched, with effects stronger in slower readers.

Blocking inner speech impairs reading comprehension but not listening comprehension. The effect is stronger for less proficient readers, more difficult texts, and second-language readers. A neuroimaging study found that direct speech in fictional text activates voice-selective auditory cortex more than indirect speech (Yao et al., 2011), suggesting readers simulate characters' voices.

Inner speech aids comprehension through multiple mechanisms: providing working memory support for maintaining sentence structure while processing, enabling integration of past concepts with current information through acoustic rehearsal, preserving prosody and intonation for interpretation, and allowing simulation of character voices. In reading development, the phonological loop is critical for learning to associate visual symbols with sounds, "sounding out" words, and building orthographic-phonological connections. Impaired phonological loop function associates with developmental dyslexia, though individuals with phonological deficits can develop compensatory strategies using visual memory and semantic approaches.

Interestingly, second-language reading shows progression from overt subvocalization and literal translation to fluent reading with minimal inner speech as proficiency develops (Kato, 2009). Skilled readers modulate inner speech based on text difficulty, reading purpose, and genre. While speed reading programs often claim to eliminate subvocalization, research suggests complete elimination is impossible and that some subvocalization aids integration and deep processing of complex material.

Forms of thinking extend beyond language

Inner speech is not the only mode of thinking. Descriptive Experience Sampling (DES) (Hurlburt et al., 2013) - a method where participants are randomly signaled throughout the day to report their current inner experience - reveals five frequent phenomena: inner speech (~26%), inner seeing/visual imagery (~34%), unsymbolized thinking (~22%), feelings (~26%), and sensory awareness (~22%). Individual variability is massive: inner speech frequency ranged from 0% to 75% across individuals, with about 25% of participants reporting no inner speech at all.

Unsymbolized thinking refers to "the experience of an explicit, differentiated thought that does not include the experience of words, images, or any other symbols." These thoughts are directly present in consciousness without accompanying words or images, yet are fully articulate and meaningful. People can readily verbalize them afterward, but do not experience them verbally in the moment. This phenomenon challenges the assumption that all thinking must involve words or images.

Visual-spatial thinking represents another major mode. Research suggests approximately 30% of people strongly use visual/spatial thinking, 45% use both visual and verbal modes, and 25% think predominantly in words. Brain imaging studies show that visual thinkers have greater activation in primary visual cortex during spontaneous thinking, while verbal thinkers show more frontal language area activation. Functional distinctions emerge: visual thinking excels at spatial reasoning, pattern recognition, and holistic processing, while verbal thinking excels at sequential reasoning, analytical tasks, and explicit rule-following.

An interesting asymmetry exists: when people engage in verbal thinking, they often involuntarily generate visual imagery (Pexman et al., 2017), but visual thinking does not necessarily evoke inner speech. This suggests visual imagery may be more fundamental or automatic than verbal representation.

Debates continue about whether the phonological loop and inner speech are identical (consensus: overlapping but not reducible to each other) and about the "true" frequency of inner speech (methodology-dependent: DES yields ~26%, questionnaires yield 60-80%+, likely capturing different aspects). The emerging view is that inner speech capacity is universal across humans, but actual use and phenomenology are highly variable across individuals and contexts.

Development follows Vygotskian progression but with substantial individual variation

Lev Vygotsky (1896-1934) was a Russian psychologist whose 1934 work "Thought and Language" laid the foundation for understanding inner speech development. He proposed that inner speech emerges through the internalization of social dialogue (Vygotsky, 1934) (Vygotsky, 1987). This view contrasted sharply with behaviorist John Watson's simpler claim that inner speech merely results from gradually quieter overt speech (Watson, 1913). Vygotsky argued that inner speech undergoes profound transformation during internalization - a process far more complex than simply becoming quieter (Alderson-Day & Fernyhough, 2015).

Vygotsky's developmental sequence has three stages. First, social speech (age 2+) is used for external communication with others. Second, private speech (age 3+) - audible self-directed talk - helps children guide their own thinking and actions. Third, inner speech (age 7+) emerges when private speech becomes fully internal and transforms into silent thought used for planning, reasoning, and abstract thinking (Vygotsky, 1978) (Berk, 1992). Vygotsky placed private speech at the center of his theory as the critical transitional process between speaking with others and thinking for oneself (Alderson-Day et al., 2020). Crucially, he argued that language doesn't just express thought but actually forms it - thought comes into existence through words rather than merely being expressed by them (Vygotsky, 1987).

Empirical evidence confirms the basic sequence

Research from 2020-2025 has largely confirmed Vygotsky's developmental sequence while revealing important nuances. Private speech - the audible self-directed talk children use during activities - typically emerges around ages 2-3 and peaks between ages 4-7, with incidence rates highest around age 5. During this peak period, children use private speech for self-regulation, planning, and task guidance. Frequency increases when children face cognitive challenges.

Internalization transforms speech into thought

The internalization process begins around age 7, following a progression from audible speech to whispers to inaudible muttering to fully silent inner speech. Phonological similarity effects - a hallmark of verbal rehearsal in working memory - were traditionally thought to emerge around age 7. However, recent studies found evidence for these effects earlier (Jarrold & Citroen, 2013), suggesting age-related differences may reflect gradual improvements in recall capacity rather than a qualitative strategy shift. Some evidence indicates the phonological loop functions as a language-learning tool from as early as 18 months.

Private speech shows clear functional significance. Task-relevant private speech correlates with better task performance, and private speech frequency increases with task difficulty. The relationship appears bidirectional: private speech aids performance, but effective use also requires sufficient attentional control. Studies using the Tower of London planning task found correlations between private speech use and phonological similarity effects, suggesting domain-general verbal mediation.

Adult inner speech shows massive individual variation

In adulthood, inner speech becomes increasingly sophisticated and may compensate for cognitive aging. Computational modeling predicted that inner speech contribution increases across the lifespan (Granato et al., 2022) in neurotypical individuals to offset declining executive functions.

However, massive individual differences exist. Descriptive Experience Sampling studies found inner speech occurred in only 20-26% of randomly sampled moments on average, with a range from 0% to 75% across individuals. The recent recognition of anendophasia - the absence or minimal experience of inner speech - challenged assumptions about inner speech universality (Nedergaard & Lupyan, 2024).

Adult inner speech also varies phenomenologically. A survey found that 77% of adults report dialogic inner speech (internal conversations), 82.5% report evaluative/motivational inner speech, 36.1% report condensed inner speech, and 25.8% report other people's voices in their inner speech (McCarthy-Jones & Fernyhough, 2011).

Neural substrates reveal dual mechanisms and surprising motor involvement

The past decade has fundamentally revised our understanding of inner speech's neural architecture. While earlier work established the involvement of Broca's area (left inferior frontal gyrus, BA 44/45) and Wernicke's area (left superior temporal gyrus, BA 22), recent research reveals two distinct neural mechanisms operating in parallel rather than in sequence.

A comprehensive meta-analysis using activation likelihood estimation (ALE) - a method for synthesizing neuroimaging data across studies - analyzed inner speech to distinguish different types (Pratts et al., 2023). The analysis identified a dual-mechanism framework. First, a corollary discharge mechanism relies on speech production regions (frontal motor areas). Second, a perceptual simulation mechanism engages speech perception regions (temporal auditory areas). Critically, both mechanisms activate concurrently during inner speech. The balance between them depends on phenomenological characteristics: whether the inner speech is deliberate versus spontaneous, and whether it uses one's own voice versus another person's voice.

Motor cortex represents inner speech more robustly than expected

The most surprising recent discovery came from intracortical recordings in ventral premotor cortex (area 6v) and area 55b in participants with paralysis (Kunz et al., 2025). Researchers achieved real-time decoding of imagined sentences from motor cortex, challenging traditional models that emphasized purely linguistic regions. Multi-unit recordings showed robust inner speech representation in motor cortex. For 50-word vocabularies, accuracy rates reached 14-33% word error rate. For 125,000-word vocabularies - approaching everyday language use - accuracy reached 26-54% word error rate. Participants preferred using inner speech over attempted speech for the brain-computer interface because it required less effort.

This motor cortex involvement raises fundamental questions about the relationship between inner speech and articulation. A complementary 2024 study recorded single neurons in the supramarginal gyrus - part of the inferior parietal lobule - and found that the same neurons represent inner, produced, and perceived speech (Wandelt et al., 2024). This suggests shared neural codes across speech modalities.

Core language areas activate concurrently, not sequentially

Frontal motor planning regions (Broca's area) may show early activation during deliberate inner speech production, while Wernicke's area simultaneously activates as part of an internal monitoring loop. The arcuate fasciculus - a major white matter tract - enables bidirectional communication between these regions. fMRI studies consistently show concurrent activation of both frontal and temporal regions, though the precise millisecond-by-millisecond dynamics remain debated. MEG studies suggested some temporal ordering (Tian & Poeppel, 2010) (Tian & Poeppel, 2013), but the relationship appears more interactive than sequential.

Beyond these core language areas, inner speech consistently recruits additional regions. The supplementary motor area (SMA and pre-SMA) supports motor sequencing and cognitive control. The anterior insula supports articulatory aspects. The inferior parietal lobule, particularly the supramarginal gyrus, supports sensorimotor integration.

When inner speech takes dialogic form - internal conversations with oneself or imagined others - additional regions activate. These include bilateral superior temporal gyri, precuneus and posterior cingulate (for self-referential processing), and right posterior superior temporal gyrus/temporoparietal junction (theory-of-mind network). Dialogic inner speech recruits social cognition networks (Alderson-Day et al., 2016), requiring representation of multiple perspectives beyond simple voice production.

The default mode network debate remains unresolved

The role of the default mode network (DMN) in inner speech remains contentious. In 2023, neuroscientist Vinod Menon proposed that the DMN integrates memory, language, and semantic representations to create an "ongoing internal narrative" central to sense of self, with potential origins in childhood self-directed speech through Vygotskian internalization (Menon, 2023). This would position the DMN as the neural substrate for our continuous inner dialogue, broadcasting integrated representations to create subjective continuity.

However, neurolinguist David Kemmerer published a critical review (Kemmerer, 2025) identifying five major challenges to this hypothesis:

1. Inner speech doesn't actually originate from self-directed overt speech
2. Massive individual differences mean any narrative isn't ongoing for everyone
3. Rodents and primates have DMN but lack language
4. Inner speech often has condensed rather than narrative form
5. Only a couple neuroscientific studies support DMN engagement during inner speech

Further complicating Menon's model, a study analyzing 1,717 participants found that verbal-predominant thought profiles were associated with segregation of language networks rather than DMN integration (Cremona et al., 2025).

An alternative view suggests the DMN contributes to semantic processing by coordinating activity across cortical regions to create embodied situation models (Fernandino & Binder, 2024), which may support inner speech without requiring it. The DMN integrates information over slower timescales than other brain systems, with activation patterns persisting during narrative processing and transitioning at event boundaries.

The emerging consensus suggests DMN involvement may be specific to spontaneous rather than deliberate inner speech, relating more to mind-wandering and self-referential thought than to volitional verbal thinking. Clinically, the DMN drives inner voice activity during unfocused states. Meditation helps regulate this network, while emerging neuromodulation techniques using transcranial focused ultrasound can target the posterior cingulate cortex to modulate DMN connectivity and alter subjective self-experience (Lord et al., 2024).

Brain-computer interfaces distinguish inner speech from attempted speech

A critical distinction has emerged in speech neuroprosthetics: most current systems target attempted speech rather than pure inner speech. Attempted speech involves motor commands for speech production, including silent or mimed attempts. Pure inner speech is imagined speech without any motor output. This distinction fundamentally impacts what BCIs can decode and their clinical applications.

The Kunz study clarified that attempted and inner speech evoke similar patterns of neural activity, but attempted speech produces stronger signals (Kunz et al., 2025). They exist on a continuum differentiated by signal magnitude rather than orthogonal subspaces. Attempted speech activates speech motor cortex even in paralyzed patients unable to produce intelligible sounds, generating robust neural signals. Inner speech - pure imagination without attempted motor output - produces similar but weaker activity patterns without subvocalization or articulation attempts.

BCI performance metrics show rapid progress

Current speech BCIs achieve remarkable performance with attempted speech. A 2023 Nature paper reported 9.1% word error rate for a 50-word vocabulary and 23.8% word error rate for a 125,000-word vocabulary - the first large-vocabulary demonstration - with speeds reaching 62 words per minute, 3.4 times the previous record (Willett et al., 2023). The system used 256 intracortical electrodes in motor cortex of a participant with ALS and unintelligible speech. A 2024 NEJM paper showed 99.6% accuracy for 50 words after just 25 days (Card et al., 2024) and 97.5% accuracy for self-paced conversation after 8.4 months of use.

For inner speech decoding, the Kunz study achieved 14-33% word error rate for 50-word vocabularies and 26-54% word error rate for 125,000-word vocabularies (Kunz et al., 2025). While lower accuracy than attempted speech, this remains functional for real-time communication. Critically, participants preferred using inner speech because it required less effort and was more comfortable than attempting to speak. International progress includes a December 2024 Chinese system (NeuroXess) achieving 71% accuracy for Chinese syllables with less than 100ms latency per character.

Key breakthroughs from 2023-2025 include large vocabulary decoding (first 125,000-word vocabulary), real-time inner speech decoding, privacy protection with password-protected decoding (over 98% accuracy), and streaming synthesis producing near-real-time audio with prosody. These advances used intracortical arrays - the gold standard providing highest spatial and temporal resolution but requiring neurosurgery. The ventral premotor cortex (area 6v) showed best performance, surprisingly outperforming Broca's area.

Clinical and atypical populations reveal functional dissociations

Studies of clinical and neuroatypical populations provide crucial insights into inner speech's functional architecture. By examining which cognitive processes remain intact and which become impaired, these dissociations reveal the boundaries of inner speech's role in cognition.

Autism shows reduced spontaneous inner speech despite intact phonological mechanisms

Individuals with autism spectrum disorder show reduced spontaneous use of inner speech for self-regulation and planning, despite intact phonological loop function. A pivotal study found that articulatory suppression impaired planning in neurotypical adults but not in autistic adults, suggesting the latter weren't spontaneously using inner speech for this task (Wallace et al., 2009). This finding was replicated (Williams et al., 2012). However, when inner speech is experimentally elicited, autistic individuals with verbal mental age above 7 years show phonological similarity effects comparable to matched controls.

The pattern is complex. Autistic children with intellectual disabilities use private speech during tasks at high rates (70%) (Niu et al., 2025), with stronger autism characteristics actually associated with more frequent private speech - possibly reflecting a compensatory strategy. Computational modeling suggests that reduced inner speech contribution creates difficulties in both early development and aging for autistic individuals.

Inner speech moderates the relationship between autism traits and emotion regulation (Albein-Urios et al., 2021). A 2025 pilot randomized controlled trial showed promise for "Thinking in Speech" therapy to improve self-regulation in autistic children (Baumann et al., 2025).

ADHD presents with delayed inner speech internalization

Children with ADHD use more private speech than typically developing peers, but with a higher proportion of task-irrelevant speech and less internalized speech. They show delays in speech internalization. The pattern suggests bidirectional relationships: unmanageable attention prevents private speech from gaining efficient behavioral control over behavior, while simultaneously, less mature inner speech impairs self-regulation.

Developmental language disorder shows delays in verbal self-regulation

Developmental language disorder (formerly specific language impairment) shows maturational delays in both onset and internalization of private speech. Children with DLD show reduced verbal mediation of cognition (Lidstone et al., 2012). Between 62-91% of children with DLD show executive function deficits by ages 3-4, with alterations across working memory, attention, processing speed, inhibition, planning, cognitive flexibility, and internalized speech. Some evidence suggests self-regulatory speech training can improve planning and problem-solving outcomes.

Environmental factors influence inner speech development. Children from cognitively and linguistically stimulating environments internalize private speech faster. Children from low verbal/social exchange environments - including Appalachian children with limited adult-child verbal communication norms and children from abuse backgrounds - showed delays (Berk & Garvin, 1984). However, fundamental mechanisms appear to transcend cultural boundaries, as demonstrated by a cross-cultural study finding consistency in private speech-phonological recoding relationships across Saudi Arabian and British samples (Al-Namlah et al., 2006).

Schizophrenia shows inner speech enhancement rather than suppression

People with schizophrenia experiencing auditory verbal hallucinations show "inner speech-induced enhancement" of N1-amplitude in auditory cortex, contrasting sharply with the normal suppression seen in healthy controls (Whitford et al., 2025). In typical individuals, inner speech suppresses N1-amplitude - an early auditory response - by dampening predicted sounds through corollary discharge mechanisms. The brain predicts the sensory consequences of inner speech and attenuates the expected input. In patients with auditory verbal hallucinations (AVH), inner speech instead enhances N1-amplitude, amplifying rather than suppressing auditory salience. The severity of AVH correlated with degree of N1-enhancement (rho=-0.222, p=0.027).

This dysfunction in predictive processing may cause inner speech to be misperceived as external voices. Beyond the neurophysiological evidence, people with schizophrenia experiencing AVH report inner speech as uncontrolled and intrusive. It overflows and distracts with more intrusions, has negative and dystonic content (self-derogatory), and is fragmented with multiple "voices" lacking coherent viewpoint (Petrolini et al., 2020). Inner speech becomes a distractor rather than a tool for self-regulation, with executive function deficits both causing and exacerbated by uncontrolled inner speech.

The misattribution hypothesis - that AVH represent inner speech misattributed as external - remains influential but oversimplifies the phenomenon. Recent models emphasize predictive processing failures and source monitoring deficits. Alternative theories include hyperactive speech perception systems generating false perceptions independent of inner speech, memory-based reactivations of previous experiences, and aberrant salience attribution to internally generated events. AVH is increasingly recognized as heterogeneous rather than arising from a single mechanism.

Treatment implications suggest targeting control and structure of inner speech rather than eliminating it. Approaches include helping patients channel inner speech toward relevant tasks, addressing fragmentation to achieve stable perspective, and using metacognitive training to improve inner speech monitoring.

Depression and anxiety reveal specific inner speech profiles

A combined EMG and self-report study (N=31) found that rumination involves different types of inner speech rather than simply more inner speech (Moffatt et al., 2020). During rumination versus distraction, participants reported 70.57% inner speech with significantly more dialogic inner speech - internal conversations (p<0.001, d=1.073) - and more evaluative inner speech - self-critical, judgmental content (p<0.001, d=0.81). Content was present- and future-oriented, focusing on current feelings and consequences. Distraction involved more visual imagery (46% versus 38%).

Surprisingly, lip muscle EMG activity showed no difference between rumination and distraction, while frontalis (brow) muscle showed more activity during rumination. This suggests inner speech differences are phenomenological rather than articulatory. The evaluative and dialogic nature potentially contributes to rumination's negative effects through verbal overshadowing - thinking verbally about problems may worsen mood compared to visual processing.

For anxiety and worry, a study (N=378) identified specific predictive inner speech types (Ghamari Kivi et al., 2023). Other-people inner speech - hearing others' voices in thought - was the strongest predictor of anxiety, depression, and somatization symptoms. Evaluative (critical) inner speech strongly predicted distress symptoms. Conversely, positive (regulatory) inner speech negatively predicted anxiety and depression, acting as a protective factor. Higher trait anxiety correlates with more self-critical inner speech. A network analysis (N=316) found voice jitter - pitch instability - shows robust negative correlation with state anxiety (Wang et al., 2025), suggesting anxiety affects speech production networks including potentially inner speech.

Aphasia reveals dissociable inner speech forms

Aphasia presents a complex and sometimes paradoxical pattern. Inner speech can be relatively preserved compared to overt speech in some individuals but more severely impaired in others. Many individuals with aphasia report experiencing inner speech despite severe output impairments (Fama & Turkeltaub, 2020). A study (N=53) found self-reported successful inner speech correlated with lexical-phonological retrieval ability and performance on phonological tasks, but not with articulatory complexity or overt naming accuracy (Fama et al., 2019).

However, other evidence shows inner speech can be more impaired than overt speech. Silent rhyming task studies found people with Broca's, conduction, and anomic aphasia showed severe impairments (Langland-Hassan et al., 2015) even when confrontation naming was relatively preserved, with some performing at chance.

A 2024 metaphor analysis examined autobiographical accounts from four individuals with aphasia (Tichborne et al., 2024), including Dr. Jill Bolte Taylor (a neuroanatomist who experienced a left-hemisphere stroke) and individuals identified as Schultz, Marks, and Broussard. The authors used systematic metaphors to describe their inner speech experiences. Phonological inner speech was described using metaphors of "words as objects" and "hearing words," while dialogic inner speech was described using metaphors of "inner voices/persons/monologue/dialogue" and "aphasia as silence."

The accounts reveal dissociable forms. Dialogic inner speech - internal conversations and "inner voices" - can be selectively impaired while phonological inner speech (sound manipulation) is preserved, as in Taylor's case. Conversely, phonological inner speech can be severely impaired while dialogic inner speech is preserved, as in Schultz's case. Schultz described "thinking without words" throughout recovery, reporting mental time-travel, planning and problem-solving from early stages. Taylor described severely impaired dialogic inner speech but retained the ability to manipulate phonological representations. The two forms have distinct neural substrates and functional properties.

Many people with moderate aphasia report experiencing inner speech more frequently after stroke compared to pre-stroke (Alexander et al., 2024). Inner speech was present in most sampled instances throughout daily life, with common themes of remembering, planning, and self-motivation. No significant relationship emerged between aphasia severity and inner speech frequency, with one important exception: people with anosognosia - lack of insight into deficits - had significantly fewer inner speech experiences, indicating connections between inner speech and self-awareness.

The critical question of whether inner speech relies primarily on Broca's area (production) or Wernicke's area (perception/phonology) remains partially answered. Both contribute, with specific impairment patterns depending on lesion location, extent, and individual compensatory strategies. The finding that inner speech can sometimes be selectively preserved after left hemisphere damage suggests right hemisphere or bilateral involvement in some individuals.

Individual differences across languages and modalities

Beyond clinical populations, remarkable variation in inner speech emerges across linguistic and sensory modalities. These differences highlight the flexibility of inner speech mechanisms and challenge assumptions about its universal form.

Multilingual speakers show complex language selection patterns

The question "what language do bilinguals think in?" proves more complex than a simple answer allows. A comprehensive study using the Bilingualism and Emotions Questionnaire surveyed 1,579 multilinguals representing 77 different L1s (first languages) and up to five languages (Dewaele & Salomidou, 2015). Results showed overall L1 preference for inner speech but not exclusively. 56.6% reported using L2 (second language) frequently or all the time for inner speech, and 37.9% used L2 frequently for mental calculations. L3, L4, and L5 showed progressively less use. However, substantial individual variation exists - some highly proficient bilinguals achieve complete language shift in inner speech.

Proficiency, age of acquisition, and social factors all contribute

Multiple interacting factors influence language choice in inner speech. Higher L2 proficiency correlates with increased L2 inner speech use, with self-perceived proficiency emerging as a significant predictor. Age of acquisition shows critical period effects: early acquirers (ages 3-7) show significantly higher rates of L2/L3 inner speech than late acquirers. Age of acquisition - not bilingualism itself - is the primary determinant of less than nativelike L2 ultimate attainment (Hyltenstam et al., 2021).

Frequency of L2 use in daily interactions emerges as the single strongest predictor of inner speech language choice, along with size of L2 social network and length of residence in L2 communities. Context-specific activation affects language selection - individuals may think in different languages depending on topic, interlocutor, and situation.

Emotional valence plays a crucial role. A systematic review found L1 preferred for emotional inner speech even when L2 is used for general inner speech, showing an "L1 advantage in emotional processing" due to increased automaticity but an "L2 advantage in emotional regulation" (Sharif & Mahmood, 2023) that allows psychological distance. Recent neuroimaging studies (2020-2023) confirm reduced L2 emotional resonance across psychophysiological measures, though naturalistic learning contexts with emotional and prosodic cues led to L2 emotional processing similar to L1 (Brase & Mani, 2017). An fMRI/ERP study found positive word processing advantage in L1 with different brain activation patterns for emotional words across languages (Chen et al., 2015).

Code-switching occurs automatically even in thought

Covert code-switching - switching between languages in inner speech - is documented in multilinguals, though less common than in overt speech.

Chinese-English bilinguals automatically translate English words into Chinese unconsciously (Zhang et al., 2011). Even with subliminal presentation of English words (59ms exposure), participants processed Chinese characters within English words. A PNAS study used brain potentials to reveal unconscious translation during foreign-language comprehension (Thierry & Wu, 2007). N400 modulation - a brain response indicating semantic processing - showed Chinese activation during English processing while participants remained unaware of hidden character repetition in Chinese translations. This automatic translation occurs even in highly proficient bilinguals at subliminal speeds.

Turning a language off requires increased engagement of the dorsolateral prefrontal cortex and anterior cingulate cortex, while turning a language on did not elicit additional activity (Baus et al., 2018). Interlocutor identity is represented in a sustained fashion throughout different stages of language planning (Fernandez et al., 2017), with language switching more common in bilingual interactional contexts and suppressed in monolingual contexts.

Language-specific experiences reshape conceptual systems

Pavlenko's analyses of autobiographical narratives revealed critical turning points in L2 inner speech development: initial L1 continuation after migration, a period of "inner speech loss" when caught between languages, L2 internalization with conceptual restructuring, loss of L1-L2 translation ability in some cases, and emergence of a new self in L2. As immigrant writer Gerda Lerner described: "It took several years before I began to think in English. It was exciting when it actually happened."

Learning L2 involves adopting new conceptual categories beyond simple translation equivalents. Thinking-for-speaking patterns differ across languages, evident in gesture studies showing conceptual shifts. The language of math instruction strongly predicts the language of mental arithmetic even decades later. Different languages evoke different emotional responses even in inner speech. Less proficient L2 learners use L2 inner speech for rehearsal and planning, while proficient speakers use it for abstract thinking. As proficiency develops, learners report a "din" phenomenon - growing mental sounds in L2.

Working memory processing differs across languages. L2 readers rely more heavily on the phonological loop (covert articulation) than L1 readers (Kato, 2009). Less proficient L2 readers show greater dependence on inner speech during reading, while advanced L2 readers develop more direct orthographic processing.

Most research focuses on bilinguals rather than multilinguals with three or more languages. How multilinguals manage language selection, competition, and switching in inner speech remains unclear. Recent fMRI studies show different neural networks activated for L1 versus L2 inner speech, with early bilinguals showing more overlapping networks than late bilinguals. Professional interpreters show unique activation patterns.

Deaf individuals experience sign-based inner speech

While research on deaf populations remains limited, available evidence indicates that deaf signers experience inner speech in visual-spatial sign language modality. Congenitally deaf ASL users report inner "speech" as "seeing" ASL signs rather than "hearing" words - a highly visual inner experience involving pictures, printed words, or signs. Those with speech therapy history report multimodal inner experience: simultaneously "hearing" voice, lipreading, and seeing signs, similar to multilingual experience.

Brain activation patterns show Broca's area processes sign language despite its visual-spatial modality. Sign language involves extensive activation of both hemispheres, unlike the predominantly left hemisphere activation for spoken language, with a highly active right frontal lobe for movement processing. Spatial awareness scores are often higher in deaf signers due to enhanced visual-spatial processing. Critically, damage to Broca's or Wernicke's areas affects signing similarly to speech, suggesting sign is brain-coded as language, not merely movement. Stuttering can even occur in sign language, parallel to spoken language, further supporting the linguistic rather than purely motoric nature of signed inner speech.

Conclusions

Inner speech research has transformed from viewing this phenomenon as a universal, monolithic process to recognizing it as a flexible cognitive tool with extraordinary individual variation across phenomenology, frequency, and form. While we have established that inner speech emerges developmentally through Vygotskian internalization, serving multiple cognitive functions from working memory to self-regulation, and recruiting both production and perception neural systems in parallel, recent discoveries - particularly robust motor cortex representation enabling real-time BCI decoding - continue to challenge traditional models. Clinical populations reveal dissociable impairments across different inner speech forms, with aphasia showing selective preservation or impairment of phonological versus dialogic inner speech, schizophrenia demonstrating predictive processing failures that enhance rather than suppress auditory responses, and autism exhibiting reduced spontaneous use despite intact mechanisms. Beyond its theoretical significance for understanding consciousness and the language-thought relationship, inner speech has emerged as a practical target for therapeutic intervention in rumination and self-regulation disorders, a biomarker for clinical conditions, and a communication channel for severely paralyzed individuals through neuroprosthetic decoding. As methodologies advance and individual differences are better characterized - including the recently discovered phenomenon of anendophasia and complex multilingual language selection patterns - inner speech research promises both fundamental insights into human cognition and tangible benefits for mental health treatment, neurological rehabilitation, and assistive technology.

References

1. Kunz, E., Willett, F., Avansino, D., Hochberg, L., Henderson, J., & Shenoy, K. (2025). Decoding inner speech for brain-computer interfaces. Cell, 188, 1-15. doi:10.1016/j.cell.2025.06.015
2. Vygotsky, L. (1934). Thought and Language. [Link]
3. Baddeley, A., & Hitch, G. (1974). Working memory. Psychology of Learning and Motivation, 8, 47-89. doi:10.1016/S0079-7421(08)60452-1
4. Oppenheim, G., & Dell, G. (2010). Motor movement matters: The flexible abstractness of inner speech. Memory & Cognition, 38(8), 1147-1160. doi:10.3758/MC.38.8.1147
5. Hughes, R. (2025). The phonological store of working memory: A critique and an alternative, perceptual-motor, approach to verbal short-term memory. Quarterly Journal of Experimental Psychology, 78(2), 240-263. doi:10.1177/17470218241257885
6. Emerson, M., & Miyake, A. (2003). The role of inner speech in task switching: A dual-task investigation. Journal of Memory and Language, 48, 148-168. doi:10.1016/S0749-596X(02)00511-9
7. Miyake, A., Emerson, M., Padilla, F., & Ahn, J. (2004). Inner speech as a retrieval aid for task goals: The effects of cue type and articulatory suppression. Acta Psychologica, 115(2-3), 123-142. doi:10.1016/j.actpsy.2003.12.004
8. Wallace, G., Silvers, J., Martin, A., & Kenworthy, L. (2009). Brief report: Further evidence for inner speech deficits in autism spectrum disorders. Journal of Autism and Developmental Disorders, 39, 1735-1739. doi:10.1007/s10803-009-0802-8
9. Senay, I., Albarracín, D., & Noguchi, K. (2010). Motivating goal-directed behavior through introspective self-talk: The role of the interrogative form of simple future tense. Psychological Science, 21(4), 499-504. doi:10.1177/0956797610364751
10. Dolcos, S., & Albarracín, D. (2014). The inner speech of behavioral regulation: Intentions and task performance strengthen when you talk to yourself as a You. European Journal of Social Psychology, 44, 636-642. doi:10.1002/ejsp.2048
11. Filik, R., & Barber, E. (2011). Inner speech during silent reading reflects the reader's regional accent. PLoS ONE, 6(10), e25782. doi:10.1371/journal.pone.0025782
12. Yao, B., Belin, P., & Scheepers, C. (2011). Silent reading of direct versus indirect speech activates voice-selective areas in the auditory cortex. Journal of Cognitive Neuroscience, 23(10), 3146-3152. doi:10.1162/jocn_a_00022
13. Kato, S. (2009). Suppressing inner speech in ESL reading: Implications for developmental changes in second language reading fluency. Modern Language Journal.
14. Hurlburt, R., Heavey, C., & Kelsey, J. (2013). Toward a phenomenology of inner speaking. Consciousness and Cognition, 22(4), 1477-1494. doi:10.1016/j.concog.2013.10.003
15. Pexman, P., Dancy, C., & McGregor, K. (2017). Verbal thinking and visual imagery. Cognition.
16. Vygotsky, L. (1987). Thinking and Speech.
17. Watson, J. (1913). Psychology as the Behaviorist Views It. Psychological Review, 20(2), 158-177. doi:10.1037/h0074428
18. Alderson-Day, B., & Fernyhough, C. (2015). Inner Speech: Development, Cognitive Functions, Phenomenology, and Neurobiology. Psychological Bulletin, 141(5), 931-965. doi:10.1037/bul0000021
19. Vygotsky, L. (1978). Mind in Society: The Development of Higher Mental Processes.
20. Berk, L. (1992). Children's Private Speech: An Overview of Theory and the Status of Research. Private Speech: From Social Interaction to Self-Regulation.
21. Alderson-Day, B., Mitrenga, K., Wilkinson, S., McCarthy-Jones, S., & Fernyhough, C. (2020). The Emergence of Inner Speech and Its Measurement in Atypically Developing Children. Frontiers in Psychology, 11, 279. doi:10.3389/fpsyg.2020.00279
22. Jarrold, C., & Citroen, R. (2013). Reevaluating key evidence for the development of rehearsal: Phonological similarity effects in children are subject to proportional scaling artifacts. Developmental Psychology, 49(5), 837-847. doi:10.1037/a0028771
23. Granato, A., Cartoni, E., Da Rold, F., Mattera, A., & Baldassarre, G. (2022). A computational model of inner speech supporting flexible goal-directed behaviour in Autism. Scientific Reports, 12, 14198. doi:10.1038/s41598-022-18445-9
24. Nedergaard, J., & Lupyan, G. (2024). Not everybody has an inner voice: Behavioral consequences of anendophasia. Psychological Science, 35(5), 506-520. doi:10.1177/09567976231225925
25. McCarthy-Jones, S., & Fernyhough, C. (2011). The varieties of inner speech: Links between quality of inner speech and psychopathological variables in a sample of young adults. Consciousness and Cognition, 20(4), 1586-1593. doi:10.1016/j.concog.2011.08.005
26. Pratts, J., Barsics, C., & Desmedt, O. (2023). Bridging phenomenology and neural mechanisms of inner speech: ALE meta-analysis on egocentricity and spontaneity in a dual-mechanistic framework. NeuroImage, 282, 120399. doi:10.1016/j.neuroimage.2023.120399
27. Wandelt, S., Bjånes, D., Pejsa, K., Lee, B., Liu, C., & Andersen, R. (2024). Representation of internal speech by single neurons in human supramarginal gyrus. Nature Human Behaviour, 8(6), 1136-1149. doi:10.1038/s41562-024-01867-y
28. Tian, X., & Poeppel, D. (2010). Mental imagery of speech and movement implicates the dynamics of internal forward models. Frontiers in Psychology.
29. Tian, X., & Poeppel, D. (2013). The effect of imagination on stimulation: The functional specificity of efference copies in speech processing. Journal of Cognitive Neuroscience.
30. Alderson-Day, B., Weis, S., McCarthy-Jones, S., Moseley, P., Smailes, D., & Fernyhough, C. (2016). The brain's conversation with itself: neural substrates of dialogic inner speech. Social Cognitive and Affective Neuroscience, 11(1), 110-120. doi:10.1093/scan/nsv094
31. Menon, V. (2023). 20 years of the default mode network: A review and synthesis. Neuron, 111(16), 2469-2487. doi:10.1016/j.neuron.2023.04.023
32. Kemmerer, D. (2025). Does the Default Mode Network Mediate an Ongoing Internal Narrative? An Evaluation of Menon's (2023) Hypothesis. Journal of Cognitive Neuroscience, 37(12), 2676-2683. doi:10.1162/jocn_a_02066
33. Cremona, S., Gillig, A., Mellet, E., & Joliot, M. (2025). A dynamic framework of brain functional patterns shaped by spontaneous thoughts beyond the default mode network. Scientific Reports, 15, 28389. doi:10.1038/s41598-025-28389-0
34. Fernandino, L., & Binder, J. (2024). How Does the "Default Mode" Network Contribute to Semantic Cognition?. Brain and Language, 252, 105405. doi:10.1016/j.bandl.2024.105405
35. Lord, B., et al. (2024). Transcranial focused ultrasound to the posterior cingulate cortex modulates default mode network and subjective experience: An fMRI pilot study. Frontiers in Human Neuroscience, 18, 1392199. doi:10.3389/fnhum.2024.1392199
36. Willett, F., et al. (2023). A high-performance speech neuroprosthesis. Nature, 620, 1031-1036. doi:10.1038/s41586-023-06377-x
37. Card, N., et al. (2024). An accurate and rapidly calibrating speech neuroprosthesis. New England Journal of Medicine, 391(7), 609-618. doi:10.1056/NEJMoa2314132
38. Williams, D., Bowler, D., & Jarrold, C. (2012). Inner speech is used to mediate short-term memory, but not planning, among intellectually high-functioning adults with autism spectrum disorder. Development and Psychopathology, 24(1), 225-239. doi:10.1017/S0954579411000794
39. Niu, Y., Wang, X., & Zhang, Y. (2025). Private speech use in autistic children with intellectual disabilities. Frontiers in Psychology.
40. Albein-Urios, N., Yao, N., Kirkovski, M., & Enticott, P. (2021). Inner speech moderates autism traits and emotion regulation. Research in Autism Spectrum Disorders.
41. Baumann, S., Russell, A., & Gaunt, D. (2025). Thinking in Speech therapy for self-regulation in autistic children. Journal of Autism and Developmental Disorders.
42. Lidstone, J., Meins, E., & Fernyhough, C. (2012). Individual differences in children's private speech. Cognitive Development.
43. Berk, L., & Garvin, R. (1984). Development of Private Speech Among Low-Income Appalachian Children. Developmental Psychology, 20(2), 271-286. doi:10.1037/0012-1649.20.2.271
44. Al-Namlah, A., Fernyhough, C., & Meins, E. (2006). Sociocultural influences on the development of verbal mediation. Developmental Psychology.
45. Whitford, T., Oestreich, L., Ford, J., Roach, B., Loewy, R., Stuart, B., & Mathalon, D. (2025). Inner speech-induced enhancement of auditory cortex responses in schizophrenia. Schizophrenia Bulletin, 51(2), 234-248. doi:10.1093/schbul/sbab001
46. Petrolini, V., Fortier, M., & Vicente, A. (2020). The role of inner speech in executive functioning. Frontiers in Psychology.
47. Moffatt, J., Mitrenga, K., Alderson-Day, B., Moseley, P., & Fernyhough, C. (2020). Inner experience differs in rumination and distraction without a change in electromyographical correlates of inner speech. PLOS ONE, 15(9), e0238920. doi:10.1371/journal.pone.0238920
48. Ghamari Kivi, H., Rehman, I., & Chaplin, T. (2023). The varieties of inner speech and their links to emotional distress. Consciousness and Cognition.
49. Wang, X., Zhang, L., & Chen, H. (2025). Voice jitter and state anxiety. Journal of Voice.
50. Fama, M., & Turkeltaub, P. (2020). Inner speech in aphasia: Current evidence, clinical implications, and future directions. American Journal of Speech-Language Pathology.
51. Fama, M., Hayward, W., Snider, S., Friedman, R., & Turkeltaub, P. (2019). The subjective experience of inner speech in aphasia is a meaningful reflection of lexical retrieval. Journal of Speech, Language, and Hearing Research, 62(1), 106-122. doi:10.1044/2018_JSLHR-L-18-0222
52. Langland-Hassan, P., Faries, F., Richardson, M., & Dietz, A. (2015). Inner speech deficits in people with aphasia. Journal of Neurolinguistics.
53. Tichborne, B., Liu, F., & Bose, A. (2024). "The illuminating quiet": A metaphor analysis of autobiographical descriptions of inner speech in aphasia. Aphasiology, 1314-1340. doi:10.1080/02687038.2024.2423930
54. Alexander, B., Carragher, M., Worrall, L., & Rose, M. (2024). Inner speech experiences in people with aphasia. Aphasiology.
55. Dewaele, J., & Salomidou, L. (2015). Language choice in multilingualism. International Journal of Multilingualism.
56. Hyltenstam, K., Bylund, E., Abrahamsson, N., & Park, H. (2021). Dominant language replacement. Bilingualism: Language and Cognition.
57. Sharif, I., & Mahmood, K. (2023). Emotional processing in first and second language. Bilingualism: Language and Cognition.
58. Brase, J., & Mani, N. (2017). Effects of learning context on emotion word acquisition. International Journal of Bilingualism.
59. Chen, P., Lin, J., Chen, B., Lu, C., & Guo, T. (2015). Processing emotional words in two languages. Cortex.
60. Zhang, S., Morris, M., Cheng, C., & Yap, A. (2011). Heritage-culture images disrupt second-language processing. Proceedings of the National Academy of Sciences.
61. Thierry, G., & Wu, Y. (2007). Brain potentials reveal unconscious translation during foreign-language comprehension. Proceedings of the National Academy of Sciences, 104(30), 12530-12535. doi:10.1073/pnas.0609927104
62. Baus, C., Branzi, F., & Costa, A. (2018). Language control in bilingual production. Neuropsychologia.
63. Fernandez, C., Litcofsky, K., & Hell, J. (2017). Neural correlates of intra-sentential code-switching. Journal of Neurolinguistics.