One glance at a whole session
A live rendering of a (fictional) patient session - emotional arc, topics, and timestamped moments the clinician can jump to instead of re-listening to an hour of audio.
The index families
Around forty objective indices, computed deterministically from the recording and transcript - counts, durations, frequencies. Not opinions.
Timing & pacing
How the conversation breathes
- Talk-time ratio and normalized words-per-minute per speaker
- Question-to-answer latency (and its trend across treatment)
- Tempo over time - a proxy for arousal and tension
"Her answer latency after your questions dropped from 5.4s to 3.6s across three sessions."
Silence & turn-taking
What happens between the words
- Silences typed by context: after a question, mid-story, at a topic change
- Interruptions and overlaps - who yields, who presses
- Longest pause, turn lengths, monologues
"Three long silences today - two right after questions about her father."
Language markers
Word choice as signal
- First-person density and absolutist words ("always", "never") - both with published links to depression and rumination
- Past / present / future orientation; insight and mental-state vocabulary
- Disfluencies per minute; tentative vs. certain phrasing
"Absolutist word rate is down 40% since intake - alongside a shift toward future-tense talk."
Voice & prosody
The tone beneath the text
- Pitch range and variability - flattened prosody is a documented depression correlate
- Loudness, jitter/shimmer, articulation vs. speaking rate
- Vocal arousal curve across the session - the "tone plot"
"Vocal arousal spikes at 31:02 - exactly where the session's pivotal moment was marked."
Topics & people
The session's map
- Topic timeline with dwell time; recurring topics across the treatment
- Per-topic shifts: how speech changes when a specific subject is on the table
- Who the client talks about, and for how long
"When the topic turns to her mother, disfluency rises 205% above her own baseline."
Therapeutic alliance
The relationship, measured gently
- Linguistic style matching - a validated rapport measure
- Prosodic synchrony; shared laughter
- Rupture and repair moments, timestamped
"Style matching climbed 0.44 to 0.64 over a month - an objective echo of the bond you feel forming."
Therapist growth (private)
A mirror only you see
- Your talk share, open-question ratio, reflections : questions
- Silence tolerance, interruption rate, own response latency
- Elicitation effectiveness: how much clients elaborate after each question type
"Open questions drew 7.7 words of elaboration on average; closed ones drew 1. Your call what to do with that."
New ยท Thought organization
Supporting the mental status exam
- Semantic coherence between consecutive ideas - a computational proxy studied for derailment and loose associations
- Question-answer relevance and "returns to the point" (tangentiality vs. circumstantiality)
- Thought-blocking candidates: pauses after unfinished clauses, timestamped for review
- Perseveration rate, lexical diversity, speech-graph connectivity
"Two blocking candidates flagged at 12:40 and 31:02 - jump to them and judge clinically."
Where the evidence stands
We are explicit about evidence tiers - clinicians deserve the honest version.
Published, replicated signals: absolutist language and first-person density in depression; speech-rate and prosodic changes in mood disorders; language style matching as a rapport measure; semantic-coherence and speech-graph measures of disorganized thought.
Face-valid measurements: most timing, silence and topic indices are precise measurements whose clinical meaning is supplied by you - the tool tracks them, per patient, over time.
What LPS never does: no diagnoses, no thresholds, no "the client is anxious." Single sessions are noisy; trends carry the signal. Hebrew lexicons are marked beta pending validation.
Selected references
- Al-Mosaiwi, M., & Johnstone, T. (2018). In an absolute state: Elevated use of absolutist words is a marker specific to anxiety, depression, and suicidal ideation. Clinical Psychological Science, 6(4), 529โ542.
- Rude, S., Gortner, E.-M., & Pennebaker, J. (2004). Language use of depressed and depression-vulnerable college students. Cognition & Emotion, 18(8), 1121โ1133.
- Boyd, R. L., Ashokkumar, A., Seraj, S., & Pennebaker, J. W. (2022). The development and psychometric properties of LIWC-22. University of Texas at Austin.
- Bedi, G., Carrillo, F., Cecchi, G. A., et al. (2015). Automated analysis of free speech predicts psychosis onset in high-risk youths. npj Schizophrenia, 1, 15030.
- Corcoran, C. M., Carrillo, F., Fernรกndez-Slezak, D., et al. (2018). Prediction of psychosis across protocols and risk cohorts using automated language analysis. World Psychiatry, 17(1), 67โ75.
- Mota, N. B., Vasconcelos, N. A. P., Lemos, N., et al. (2012). Speech graphs provide a quantitative measure of thought disorder in psychosis. PLoS ONE, 7(4), e34928.
- Mota, N. B., Copelli, M., & Ribeiro, S. (2017). Thought disorder measured as random speech structure classifies negative symptoms and schizophrenia diagnosis 6 months in advance. npj Schizophrenia, 3, 18.
- Elvevรฅg, B., Foltz, P. W., Weinberger, D. R., & Goldberg, T. E. (2007). Quantifying incoherence in speech: An automated methodology and novel application to schizophrenia. Schizophrenia Research, 93(1โ3), 304โ316.
- Cannizzaro, M., Harel, B., Reilly, N., Chappell, P., & Snyder, P. J. (2004). Voice acoustical measurement of the severity of major depression. Brain and Cognition, 56(1), 30โ35.
- Mundt, J. C., Vogel, A. P., Feltner, D. E., & Lenderking, W. R. (2012). Vocal acoustic biomarkers of depression severity and treatment response. Biological Psychiatry, 72(7), 580โ587.
- Ireland, M. E., Slatcher, R. B., Eastwick, P. W., et al. (2011). Language style matching predicts relationship initiation and stability. Psychological Science, 22(1), 39โ44.
- Imel, Z. E., Barco, J. S., Brown, H. J., et al. (2014). The association of therapist empathy and synchrony in vocally encoded arousal. Journal of Counseling Psychology, 61(1), 146โ153.
Privacy, structurally. All processing runs inside your environment - on the device or on servers you control. There is no LPS cloud holding recordings; the architecture makes the privacy promise, not a policy document.
See your own session, measured
A 20-minute demo with a fictional session - or your own recording, on your own machine.
Request a demo